[ { "path": ".devcontainer/Dockerfile", "content": "FROM mcr.microsoft.com/devcontainers/typescript-node:1-22-bookworm\n\n# Install MongoDB tools (mongosh, mongorestore, mongodump) directly from MongoDB repository\nRUN curl -fsSL https://www.mongodb.org/static/pgp/server-8.0.asc | gpg --dearmor -o /usr/share/keyrings/mongodb-server-8.0.gpg && \\\n echo \"deb [ signed-by=/usr/share/keyrings/mongodb-server-8.0.gpg ] http://repo.mongodb.org/apt/debian bookworm/mongodb-org/8.0 main\" | tee /etc/apt/sources.list.d/mongodb-org-8.0.list && \\\n apt-get update && \\\n apt-get install -y mongodb-mongosh mongodb-database-tools vim && \\\n apt-get autoremove -y && \\\n rm -rf /var/lib/apt/lists/*\n" }, { "path": ".devcontainer/devcontainer.json", "content": "// For format details, see https://aka.ms/devcontainer.json. For config options, see the\n// README at: https://github.com/devcontainers/templates/tree/main/src/typescript-node\n{\n\t\"name\": \"Node.js & TypeScript\",\n\t// Or use a Dockerfile or Docker Compose file. More info: https://containers.dev/guide/dockerfile\n\t\"build\": {\n\t\t\"dockerfile\": \"Dockerfile\"\n\t},\n\n\t\"customizations\": {\n\t\t\"vscode\": {\n\t\t\t\"extensions\": [\"esbenp.prettier-vscode\", \"dbaeumer.vscode-eslint\", \"svelte.svelte-vscode\"]\n\t\t}\n\t},\n\n\t\"features\": {\n\t\t// Install docker in container\n\t\t\"ghcr.io/devcontainers/features/docker-in-docker:2\": {\n\t\t\t// Use proprietary docker engine. I get a timeout error when using the default moby engine and loading\n\t\t\t// microsoft's PGP keys\n\t\t\t\"moby\": false\n\t\t}\n\t}\n\n\t// Use 'forwardPorts' to make a list of ports inside the container available locally.\n\t// \"forwardPorts\": [],\n\n\t// Use 'postCreateCommand' to run commands after the container is created.\n\t// \"postCreateCommand\": \"yarn install\",\n\n\t// Configure tool-specific properties.\n\t// \"customizations\": {},\n\n\t// Uncomment to connect as root instead. More info: https://aka.ms/dev-containers-non-root.\n\t// \"remoteUser\": \"root\"\n}\n" }, { "path": ".dockerignore", "content": "Dockerfile\n.vscode/\n.idea\n.gitignore\nLICENSE\nREADME.md\nnode_modules/\n.svelte-kit/\n.env*\n!.env\n.env.local\ndb\nmodels/**" }, { "path": ".env", "content": "# Use .env.local to change these variables\n# DO NOT EDIT THIS FILE WITH SENSITIVE DATA\n\n### Models ###\n# Models are sourced exclusively from an OpenAI-compatible base URL.\n# Example: https://router.huggingface.co/v1\nOPENAI_BASE_URL=https://router.huggingface.co/v1\n\n# Canonical auth token for any OpenAI-compatible provider\nOPENAI_API_KEY=#your provider API key (works for HF router, OpenAI, LM Studio, etc.). \n# When set to true, user token will be used for inference calls\nUSE_USER_TOKEN=false\n# Automatically redirect to oauth login page if user is not logged in, when set to \"true\"\nAUTOMATIC_LOGIN=false\n\n### MongoDB ###\nMONGODB_URL=#your mongodb URL here, use chat-ui-db image if you don't want to set this\nMONGODB_DB_NAME=chat-ui\nMONGODB_DIRECT_CONNECTION=false\n\n\n## Public app configuration ##\nPUBLIC_APP_NAME=ChatUI # name used as title throughout the app\nPUBLIC_APP_ASSETS=chatui # used to find logos & favicons in static/$PUBLIC_APP_ASSETS\nPUBLIC_APP_DESCRIPTION=\"Making the community's best AI chat models available to everyone.\"# description used throughout the app\nPUBLIC_ORIGIN=\nPUBLIC_SHARE_PREFIX=\nPUBLIC_GOOGLE_ANALYTICS_ID=\nPUBLIC_PLAUSIBLE_SCRIPT_URL=\nPUBLIC_APPLE_APP_ID=\n\nCOUPLE_SESSION_WITH_COOKIE_NAME=\n# when OPEN_ID is configured, users are required to login after the welcome modal\nOPENID_CLIENT_ID=\"\" # You can set to \"__CIMD__\" for automatic oauth app creation when deployed, see https://datatracker.ietf.org/doc/draft-ietf-oauth-client-id-metadata-document/\nOPENID_CLIENT_SECRET=\nOPENID_SCOPES=\"openid profile inference-api read-mcp read-billing\"\nUSE_USER_TOKEN=\nAUTOMATIC_LOGIN=# if true authentication is required on all routes\n\n### Local Storage ###\nMONGO_STORAGE_PATH= # where is the db folder stored\n\n## Models overrides\nMODELS=\n\n## Task model\n# Optional: set to the model id/name from the `${OPENAI_BASE_URL}/models` list\n# to use for internal tasks (title summarization, etc). If not set, the current model will be used\nTASK_MODEL=\n\n# Arch router (OpenAI-compatible) endpoint base URL used for route selection\n# Example: https://api.openai.com/v1 or your hosted Arch endpoint\nLLM_ROUTER_ARCH_BASE_URL=\n\n## LLM Router Configuration\n# Path to routes policy (JSON array). Required when the router is enabled; must point to a valid JSON file.\nLLM_ROUTER_ROUTES_PATH=\n\n# Model used at the Arch router endpoint for selection\nLLM_ROUTER_ARCH_MODEL=\n\n# Fallback behavior\n# Route to map \"other\" to (must exist in routes file)\nLLM_ROUTER_OTHER_ROUTE=casual_conversation\n# Model to call if the Arch selection fails entirely\nLLM_ROUTER_FALLBACK_MODEL=\n# Arch selection timeout in milliseconds (default 10000)\nLLM_ROUTER_ARCH_TIMEOUT_MS=10000\n# Maximum length (in characters) for assistant messages sent to router for route selection (default 500)\nLLM_ROUTER_MAX_ASSISTANT_LENGTH=500\n# Maximum length (in characters) for previous user messages sent to router (latest user message not trimmed, default 400)\nLLM_ROUTER_MAX_PREV_USER_LENGTH=400\n\n# Enable router multimodal handling (set to true to allow image inputs via router)\nLLM_ROUTER_ENABLE_MULTIMODAL=\n# Required when LLM_ROUTER_ENABLE_MULTIMODAL=true: id or name of the multimodal model to use for image requests\nLLM_ROUTER_MULTIMODAL_MODEL=\n\n# Enable router tool support (set to true to allow tool calling via router)\nLLM_ROUTER_ENABLE_TOOLS=\n# Required when tools are active: id or name of the model to use for MCP tool calls.\nLLM_ROUTER_TOOLS_MODEL=\n\n# Router UI overrides (client-visible)\n# Public display name for the router entry in the model list. Defaults to \"Omni\".\nPUBLIC_LLM_ROUTER_DISPLAY_NAME=Omni\n# Optional: public logo URL for the router entry. If unset, the UI shows a Carbon icon.\nPUBLIC_LLM_ROUTER_LOGO_URL=\n# Public alias id used for the virtual router model (Omni). Defaults to \"omni\".\nPUBLIC_LLM_ROUTER_ALIAS_ID=omni\n\n### Transcription ###\n# Voice-to-text transcription using Whisper models\n# If set, enables the microphone button in the chat input\n# Example: openai/whisper-large-v3-turbo\nTRANSCRIPTION_MODEL=\n# Optional: Base URL for transcription API (defaults to HF inference)\n# Default: https://router.huggingface.co/hf-inference/models\nTRANSCRIPTION_BASE_URL=\n\n### Authentication ###\n# Parameters to enable open id login\nOPENID_CONFIG=\n# if it's defined, only these emails will be allowed to use login\nALLOWED_USER_EMAILS=[]\n# If it's defined, users with emails matching these domains will also be allowed to use login\nALLOWED_USER_DOMAINS=[]\n# valid alternative redirect URLs for OAuth, used for HuggingChat apps\nALTERNATIVE_REDIRECT_URLS=[] \n### Cookies\n# name of the cookie used to store the session\nCOOKIE_NAME=hf-chat\n# If the value of this cookie changes, the session is destroyed. Useful if chat-ui is deployed on a subpath\n# of your domain, and you want chat ui sessions to reset if the user's auth changes\nCOUPLE_SESSION_WITH_COOKIE_NAME=\n# specify secure behaviour for cookies \nCOOKIE_SAMESITE=# can be \"lax\", \"strict\", \"none\" or left empty\nCOOKIE_SECURE=# set to true to only allow cookies over https\nTRUSTED_EMAIL_HEADER=# header to use to get the user email, only use if you know what you are doing\n\n### Admin stuff ###\nADMIN_CLI_LOGIN=true # set to false to disable the CLI login\nADMIN_TOKEN=#We recommend leaving this empty, you can get the token from the terminal.\n\n### Feature Flags ###\nLLM_SUMMARIZATION=true # generate conversation titles with LLMs\n \nALLOW_IFRAME=true # Allow the app to be embedded in an iframe\n\n# Base servers list (JSON array). Example: MCP_SERVERS=[{\"name\": \"Web Search (Exa)\", \"url\": \"https://mcp.exa.ai/mcp\"}, {\"name\": \"Hugging Face\", \"url\": \"https://hf.co/mcp\"}]\nMCP_SERVERS=\n# When true, forward the logged-in user's Hugging Face access token\nMCP_FORWARD_HF_USER_TOKEN=\n# Exa API key (injected at runtime into mcp.exa.ai URLs as ?exaApiKey=)\nEXA_API_KEY=\n# Timeout in milliseconds for MCP tool calls (default: 120000 = 2 minutes)\nMCP_TOOL_TIMEOUT_MS=\nENABLE_DATA_EXPORT=true\n\n### Rate limits ### \n# See `src/lib/server/usageLimits.ts`\n# {\n# conversations: number, # how many conversations\n# messages: number, # how many messages in a conversation\n# assistants: number, # how many assistants\n# messageLength: number, # how long can a message be before we cut it off\n# messagesPerMinute: number, # how many messages per minute\n# tools: number # how many tools\n# }\nUSAGE_LIMITS={}\n\n### HuggingFace specific ###\n## Feature flag & admin settings\n# Used for setting early access & admin flags to users\nHF_ORG_ADMIN=\nHF_ORG_EARLY_ACCESS=\nWEBHOOK_URL_REPORT_ASSISTANT=#provide slack webhook url to get notified for reports/feature requests\n\n\n### Metrics ###\nMETRICS_ENABLED=false\nMETRICS_PORT=5565\nLOG_LEVEL=info\n\n\n### Parquet export ###\n# Not in use anymore but useful to export conversations to a parquet file as a HuggingFace dataset\nPARQUET_EXPORT_DATASET=\nPARQUET_EXPORT_HF_TOKEN=\nADMIN_API_SECRET=# secret to admin API calls, like computing usage stats or exporting parquet data\n\n### Config ###\nENABLE_CONFIG_MANAGER=true\n\n### Docker build variables ### \n# These values cannot be updated at runtime\n# They need to be passed when building the docker image\n# See https://github.com/huggingface/chat-ui/main/.github/workflows/deploy-prod.yml#L44-L47\nAPP_BASE=\"\" # base path of the app, e.g. /chat, left blank as default\n### Body size limit for SvelteKit https://svelte.dev/docs/kit/adapter-node#Environment-variables-BODY_SIZE_LIMIT\nBODY_SIZE_LIMIT=15728640\nPUBLIC_COMMIT_SHA=\n\n### LEGACY parameters\nALLOW_INSECURE_COOKIES=false # LEGACY! Use COOKIE_SECURE and COOKIE_SAMESITE instead\nPARQUET_EXPORT_SECRET=#DEPRECATED, use ADMIN_API_SECRET instead\nRATE_LIMIT= # /!\\ DEPRECATED definition of messages per minute. Use USAGE_LIMITS.messagesPerMinute instead\nOPENID_NAME_CLAIM=\"name\" # Change to \"username\" for some providers that do not provide name\nOPENID_PROVIDER_URL=https://huggingface.co # for Google, use https://accounts.google.com\nOPENID_TOLERANCE=\nOPENID_RESOURCE=\nEXPOSE_API=# deprecated, API is now always exposed\n" }, { "path": ".env.ci", "content": "MONGODB_URL=mongodb://localhost:27017/" }, { "path": ".eslintignore", "content": ".DS_Store\nnode_modules\n/build\n/.svelte-kit\n/package\n.env\n.env.*\n!.env.example\n\n# Ignore files for PNPM, NPM and YARN\npnpm-lock.yaml\npackage-lock.json\nyarn.lock\n" }, { "path": ".eslintrc.cjs", "content": "module.exports = {\n\troot: true,\n\tparser: \"@typescript-eslint/parser\",\n\textends: [\n\t\t\"eslint:recommended\",\n\t\t\"plugin:@typescript-eslint/recommended\",\n\t\t\"plugin:svelte/recommended\",\n\t\t\"prettier\",\n\t],\n\tplugins: [\"@typescript-eslint\"],\n\tignorePatterns: [\"*.cjs\"],\n\toverrides: [\n\t\t{\n\t\t\tfiles: [\"*.svelte\"],\n\t\t\tparser: \"svelte-eslint-parser\",\n\t\t\tparserOptions: {\n\t\t\t\tparser: \"@typescript-eslint/parser\",\n\t\t\t},\n\t\t},\n\t],\n\tparserOptions: {\n\t\tsourceType: \"module\",\n\t\tecmaVersion: 2020,\n\t\textraFileExtensions: [\".svelte\"],\n\t},\n\trules: {\n\t\t\"no-empty\": \"off\",\n\t\t\"require-yield\": \"off\",\n\t\t\"@typescript-eslint/no-explicit-any\": \"error\",\n\t\t\"@typescript-eslint/no-non-null-assertion\": \"error\",\n\t\t\"@typescript-eslint/no-unused-vars\": [\n\t\t\t// prevent variables with a _ prefix from being marked as unused\n\t\t\t\"error\",\n\t\t\t{\n\t\t\t\targsIgnorePattern: \"^_\",\n\t\t\t},\n\t\t],\n\t\t\"object-shorthand\": [\"error\", \"always\"],\n\t},\n\tenv: {\n\t\tbrowser: true,\n\t\tes2017: true,\n\t\tnode: true,\n\t},\n};\n" }, { "path": ".github/ISSUE_TEMPLATE/bug-report--chat-ui-.md", "content": "---\nname: Bug Report (chat-ui)\nabout: Use this for confirmed issues with chat-ui\ntitle: \"\"\nlabels: bug\nassignees: \"\"\n---\n\n## Bug description\n\n\n\n## Steps to reproduce\n\n\n\n## Screenshots\n\n\n\n## Context\n\n### Logs\n\n\n\n```\n// logs here if relevant\n```\n\n### Specs\n\n- **OS**:\n- **Browser**:\n- **chat-ui commit**:\n\n### Config\n\n\n\n## Notes\n\n\n" }, { "path": ".github/ISSUE_TEMPLATE/config-support.md", "content": "---\nname: Config Support\nabout: Help with setting up chat-ui locally\ntitle: \"\"\nlabels: support\nassignees: \"\"\n---\n\n**Please use the discussions on GitHub** for getting help with setting things up instead of opening an issue: https://github.com/huggingface/chat-ui/discussions\n" }, { "path": ".github/ISSUE_TEMPLATE/feature-request--chat-ui-.md", "content": "---\nname: Feature Request (chat-ui)\nabout: Suggest new features to be added to chat-ui\ntitle: \"\"\nlabels: enhancement\nassignees: \"\"\n---\n\n## Describe your feature request\n\n\n\n## Screenshots (if relevant)\n\n## Implementation idea\n\n\n" }, { "path": ".github/ISSUE_TEMPLATE/huggingchat.md", "content": "---\nname: HuggingChat\nabout: Requests & reporting outages on HuggingChat, the hosted version of chat-ui.\ntitle: \"\"\nlabels: huggingchat\nassignees: \"\"\n---\n\n**Do not use GitHub issues** for requesting models on HuggingChat or reporting issues with HuggingChat being down/overloaded.\n\n**Use the discussions page on the hub instead:** https://huggingface.co/spaces/huggingchat/chat-ui/discussions\n" }, { "path": ".github/release.yml", "content": "changelog:\n exclude:\n labels:\n - huggingchat\n - CI/CD\n - documentation\n categories:\n - title: Features\n labels:\n - enhancement\n - title: Bugfixes\n labels:\n - bug\n - title: Other changes\n labels:\n - \"*\"\n" }, { "path": ".github/workflows/build-docs.yml", "content": "name: Build documentation\n\non:\n push:\n branches:\n - main\n - v*-release\n\njobs:\n build:\n uses: huggingface/doc-builder/.github/workflows/build_main_documentation.yml@main\n with:\n commit_sha: ${{ github.sha }}\n package: chat-ui\n additional_args: --not_python_module\n secrets:\n token: ${{ secrets.HUGGINGFACE_PUSH }}\n hf_token: ${{ secrets.HF_DOC_BUILD_PUSH }}\n" }, { "path": ".github/workflows/build-image.yml", "content": "name: Build and Publish Image\n\npermissions:\n packages: write\n\non:\n push:\n branches:\n - \"main\"\n pull_request:\n branches:\n - \"*\"\n paths:\n - \"Dockerfile\"\n - \"entrypoint.sh\"\n workflow_dispatch:\n release:\n types: [published, edited]\n\njobs:\n build-and-publish-image-with-db:\n runs-on:\n group: aws-general-8-plus\n steps:\n - name: Checkout\n uses: actions/checkout@v4\n\n - name: Extract package version\n id: package-version\n run: |\n VERSION=$(jq -r .version package.json)\n echo \"VERSION=$VERSION\" >> $GITHUB_OUTPUT\n MAJOR=$(echo $VERSION | cut -d '.' -f1)\n echo \"MAJOR=$MAJOR\" >> $GITHUB_OUTPUT\n MINOR=$(echo $VERSION | cut -d '.' -f1).$(echo $VERSION | cut -d '.' -f2)\n echo \"MINOR=$MINOR\" >> $GITHUB_OUTPUT\n\n - name: Docker metadata\n id: meta\n uses: docker/metadata-action@v5\n with:\n images: |\n ghcr.io/huggingface/chat-ui-db\n tags: |\n type=raw,value=${{ steps.package-version.outputs.VERSION }},enable=${{github.event_name == 'release'}}\n type=raw,value=${{ steps.package-version.outputs.MAJOR }},enable=${{github.event_name == 'release'}}\n type=raw,value=${{ steps.package-version.outputs.MINOR }},enable=${{github.event_name == 'release'}}\n type=raw,value=latest,enable={{is_default_branch}}\n type=sha,enable={{is_default_branch}}\n\n - name: Set up QEMU\n uses: docker/setup-qemu-action@v3\n\n - name: Set up Docker Buildx\n uses: docker/setup-buildx-action@v3\n\n - name: Login to GitHub Container Registry\n if: github.event_name != 'pull_request'\n uses: docker/login-action@v3\n with:\n registry: ghcr.io\n username: ${{ github.repository_owner }}\n password: ${{ secrets.GITHUB_TOKEN }}\n\n - name: Inject slug/short variables\n uses: rlespinasse/github-slug-action@v4.5.0\n\n - name: Build and Publish Docker Image with DB\n uses: docker/build-push-action@v5\n with:\n context: .\n file: Dockerfile\n push: ${{ github.event_name != 'pull_request' }}\n tags: ${{ steps.meta.outputs.tags }}\n labels: ${{ steps.meta.outputs.labels }}\n platforms: linux/amd64,linux/arm64\n cache-from: type=gha\n cache-to: type=gha,mode=max\n build-args: |\n INCLUDE_DB=true\n PUBLIC_COMMIT_SHA=${{ env.GITHUB_SHA_SHORT }}\n build-and-publish-image-nodb:\n runs-on:\n group: aws-general-8-plus\n steps:\n - name: Checkout\n uses: actions/checkout@v4\n\n - name: Extract package version\n id: package-version\n run: |\n VERSION=$(jq -r .version package.json)\n echo \"VERSION=$VERSION\" >> $GITHUB_OUTPUT\n MAJOR=$(echo $VERSION | cut -d '.' -f1)\n echo \"MAJOR=$MAJOR\" >> $GITHUB_OUTPUT\n MINOR=$(echo $VERSION | cut -d '.' -f1).$(echo $VERSION | cut -d '.' -f2)\n echo \"MINOR=$MINOR\" >> $GITHUB_OUTPUT\n\n - name: Docker metadata\n id: meta\n uses: docker/metadata-action@v5\n with:\n images: |\n ghcr.io/huggingface/chat-ui\n tags: |\n type=raw,value=${{ steps.package-version.outputs.VERSION }},enable=${{github.event_name == 'release'}}\n type=raw,value=${{ steps.package-version.outputs.MAJOR }},enable=${{github.event_name == 'release'}}\n type=raw,value=${{ steps.package-version.outputs.MINOR }},enable=${{github.event_name == 'release'}}\n type=raw,value=latest,enable={{is_default_branch}}\n type=sha,enable={{is_default_branch}}\n\n - name: Set up QEMU\n uses: docker/setup-qemu-action@v3\n\n - name: Set up Docker Buildx\n uses: docker/setup-buildx-action@v3\n\n - name: Login to GitHub Container Registry\n if: github.event_name != 'pull_request'\n uses: docker/login-action@v3\n with:\n registry: ghcr.io\n username: ${{ github.repository_owner }}\n password: ${{ secrets.GITHUB_TOKEN }}\n\n - name: Inject slug/short variables\n uses: rlespinasse/github-slug-action@v4.5.0\n\n - name: Build and Publish Docker Image without DB\n uses: docker/build-push-action@v5\n with:\n context: .\n file: Dockerfile\n push: ${{ github.event_name != 'pull_request' }}\n tags: ${{ steps.meta.outputs.tags }}\n labels: ${{ steps.meta.outputs.labels }}\n platforms: linux/amd64,linux/arm64\n cache-from: type=gha\n cache-to: type=gha,mode=max\n build-args: |\n INCLUDE_DB=false\n PUBLIC_COMMIT_SHA=${{ env.GITHUB_SHA_SHORT }}\n" }, { "path": ".github/workflows/build-pr-docs.yml", "content": "name: Build PR Documentation\n\non:\n pull_request:\n paths:\n - \"docs/source/**\"\n - \".github/workflows/build-pr-docs.yml\"\n\nconcurrency:\n group: ${{ github.workflow }}-${{ github.head_ref || github.run_id }}\n cancel-in-progress: true\n\njobs:\n build:\n uses: huggingface/doc-builder/.github/workflows/build_pr_documentation.yml@main\n with:\n commit_sha: ${{ github.event.pull_request.head.sha }}\n pr_number: ${{ github.event.number }}\n package: chat-ui\n additional_args: --not_python_module\n" }, { "path": ".github/workflows/deploy-dev.yml", "content": "name: Deploy to ephemeral\non:\n pull_request:\n types: [opened, reopened, synchronize, labeled, unlabeled]\n\njobs:\n branch-slug:\n uses: ./.github/workflows/slugify.yaml\n with:\n value: ${{ github.head_ref }}\n\n deploy-dev:\n if: contains(github.event.pull_request.labels.*.name, 'preview')\n runs-on: ubuntu-latest\n needs: branch-slug\n environment:\n name: dev\n url: https://${{ needs.branch-slug.outputs.slug }}.chat-dev.huggingface.tech/chat/\n steps:\n - name: Checkout\n uses: actions/checkout@v4\n\n - name: Login to Registry\n uses: docker/login-action@v3\n with:\n username: ${{ secrets.DOCKERHUB_USERNAME }}\n password: ${{ secrets.DOCKERHUB_PASSWORD }}\n\n - name: Inject slug/short variables\n uses: rlespinasse/github-slug-action@v4.5.0\n\n - name: Set GITHUB_SHA_SHORT from PR\n if: env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT != null\n run: echo \"GITHUB_SHA_SHORT=${{ env.GITHUB_EVENT_PULL_REQUEST_HEAD_SHA_SHORT }}\" >> $GITHUB_ENV\n\n - name: Docker metadata\n id: meta\n uses: docker/metadata-action@v5\n with:\n images: |\n huggingface/chat-ui\n tags: |\n type=raw,value=dev-${{ env.GITHUB_SHA_SHORT }}\n\n - name: Set up Docker Buildx\n uses: docker/setup-buildx-action@v3\n\n - name: Build and Publish HuggingChat image\n uses: docker/build-push-action@v5\n with:\n context: .\n file: Dockerfile\n push: true\n tags: ${{ steps.meta.outputs.tags }}\n labels: ${{ steps.meta.outputs.labels }}\n platforms: linux/amd64\n cache-to: type=gha,mode=max,scope=amd64\n cache-from: type=gha,scope=amd64\n provenance: false\n build-args: |\n INCLUDE_DB=false\n APP_BASE=/chat\n PUBLIC_COMMIT_SHA=${{ env.GITHUB_SHA_SHORT }}\n" }, { "path": ".github/workflows/deploy-prod.yml", "content": "name: Deploy to k8s\non:\n # run this workflow manually from the Actions tab\n workflow_dispatch:\n\njobs:\n build-and-publish-huggingchat-image:\n runs-on: ubuntu-latest\n steps:\n - name: Checkout\n uses: actions/checkout@v4\n\n - name: Login to Registry\n uses: docker/login-action@v3\n with:\n username: ${{ secrets.DOCKERHUB_USERNAME }}\n password: ${{ secrets.DOCKERHUB_PASSWORD }}\n\n - name: Docker metadata\n id: meta\n uses: docker/metadata-action@v5\n with:\n images: |\n huggingface/chat-ui\n tags: |\n type=raw,value=latest,enable={{is_default_branch}}\n type=sha,enable=true,prefix=sha-,format=short,sha-len=8\n\n - name: Set up Docker Buildx\n uses: docker/setup-buildx-action@v3\n\n - name: Inject slug/short variables\n uses: rlespinasse/github-slug-action@v4.5.0\n\n - name: Build and Publish HuggingChat image\n uses: docker/build-push-action@v5\n with:\n context: .\n file: Dockerfile\n push: ${{ github.event_name != 'pull_request' }}\n tags: ${{ steps.meta.outputs.tags }}\n labels: ${{ steps.meta.outputs.labels }}\n platforms: linux/amd64\n cache-to: type=gha,mode=max,scope=amd64\n cache-from: type=gha,scope=amd64\n provenance: false\n build-args: |\n INCLUDE_DB=false\n APP_BASE=/chat\n PUBLIC_COMMIT_SHA=${{ env.GITHUB_SHA_SHORT }}\n deploy:\n name: Deploy on prod\n runs-on: ubuntu-latest\n needs: [\"build-and-publish-huggingchat-image\"]\n steps:\n - name: Inject slug/short variables\n uses: rlespinasse/github-slug-action@v4.5.0\n\n - name: Gen values\n run: |\n VALUES=$(cat <<-END\n image:\n tag: \"sha-${{ env.GITHUB_SHA_SHORT }}\"\n END\n )\n echo \"VALUES=$(echo \"$VALUES\" | yq -o=json | jq tostring)\" >> $GITHUB_ENV\n\n - name: Deploy on infra-deployments\n uses: aurelien-baudet/workflow-dispatch@v2\n with:\n workflow: Update application single value\n repo: huggingface/infra-deployments\n wait-for-completion: true\n wait-for-completion-interval: 10s\n display-workflow-run-url-interval: 10s\n ref: refs/heads/main\n token: ${{ secrets.GIT_TOKEN_INFRA_DEPLOYMENT }}\n inputs: '{\"path\": \"hub/chat-ui/chat-ui.yaml\", \"value\": ${{ env.VALUES }}, \"url\": \"${{ github.event.head_commit.url }}\"}'\n" }, { "path": ".github/workflows/lint-and-test.yml", "content": "name: Lint and test\n\non:\n pull_request:\n push:\n branches:\n - main\n\njobs:\n lint:\n runs-on: ubuntu-latest\n timeout-minutes: 10\n\n steps:\n - uses: actions/checkout@v3\n\n - uses: actions/setup-node@v3\n with:\n node-version: \"20\"\n cache: \"npm\"\n - run: |\n npm install ci\n - name: \"Checking lint/format errors\"\n run: |\n npm run lint\n - name: \"Checking type errors\"\n run: |\n npm run check\n\n test:\n runs-on: ubuntu-latest\n timeout-minutes: 10\n\n steps:\n - uses: actions/checkout@v3\n - uses: actions/setup-node@v3\n with:\n node-version: \"20\"\n cache: \"npm\"\n - run: |\n npm ci\n npx playwright install\n - name: \"Tests\"\n run: |\n npm run test\n\n build-check:\n runs-on:\n group: aws-general-8-plus\n timeout-minutes: 10\n steps:\n - uses: actions/checkout@v3\n - name: Build Docker image\n run: |\n docker build \\\n --build-arg INCLUDE_DB=true \\\n -t chat-ui-test:latest .\n\n - name: Run Docker container\n run: |\n export DOTENV_LOCAL=$(<.env.ci)\n docker run -d --rm --network=host \\\n --name chat-ui-test \\\n -e DOTENV_LOCAL=\"$DOTENV_LOCAL\" \\\n chat-ui-test:latest\n\n - name: Wait for server to start\n run: |\n for i in {1..10}; do\n if curl -s -o /dev/null -w \"%{http_code}\" http://localhost:3000/ | grep -q \"200\"; then\n echo \"Server is up\"\n exit 0\n fi\n echo \"Waiting for server...\"\n sleep 2\n done\n echo \"Server did not start in time\"\n docker logs chat-ui-test\n exit 1\n\n - name: Stop Docker container\n if: always()\n run: |\n docker stop chat-ui-test || true\n" }, { "path": ".github/workflows/slugify.yaml", "content": "name: Generate Branch Slug\n\non:\n workflow_call:\n inputs:\n value:\n description: \"Value to slugify\"\n required: true\n type: string\n outputs:\n slug:\n description: \"Slugified value\"\n value: ${{ jobs.generate-slug.outputs.slug }}\n\njobs:\n generate-slug:\n runs-on: ubuntu-latest\n outputs:\n slug: ${{ steps.slugify.outputs.slug }}\n\n steps:\n - name: Setup Go\n uses: actions/setup-go@v5\n with:\n go-version: \"1.21\"\n\n - name: Generate slug\n id: slugify\n run: |\n # Create working directory\n mkdir -p $HOME/slugify\n cd $HOME/slugify\n\n # Create Go script\n cat > main.go << 'EOF'\n package main\n\n import (\n \"fmt\"\n \"os\"\n \"github.com/gosimple/slug\"\n )\n\n func main() {\n if len(os.Args) < 2 {\n fmt.Println(\"Usage: slugify \")\n os.Exit(1)\n }\n\n text := os.Args[1]\n slugged := slug.Make(text)\n fmt.Println(slugged)\n }\n EOF\n\n # Initialize module and install dependency\n go mod init slugify\n go mod tidy\n go get github.com/gosimple/slug\n\n # Build\n go build -o slugify main.go\n\n # Generate slug\n VALUE=\"${{ inputs.value }}\"\n echo \"Input value: $VALUE\"\n\n SLUG=$(./slugify \"$VALUE\")\n echo \"Generated slug: $SLUG\"\n\n # Export\n echo \"slug=$SLUG\" >> $GITHUB_OUTPUT\n" }, { "path": ".github/workflows/trufflehog.yml", "content": "on:\n push:\n\nname: Secret Leaks\n\njobs:\n trufflehog:\n runs-on: ubuntu-latest\n steps:\n - name: Checkout code\n uses: actions/checkout@v4\n with:\n fetch-depth: 0\n - name: Secret Scanning\n uses: trufflesecurity/trufflehog@main\n with:\n extra_args: --results=verified,unknown\n" }, { "path": ".github/workflows/upload-pr-documentation.yml", "content": "name: Upload PR Documentation\n\non:\n workflow_run:\n workflows: [\"Build PR Documentation\"]\n types:\n - completed\n\njobs:\n build:\n uses: huggingface/doc-builder/.github/workflows/upload_pr_documentation.yml@main\n with:\n package_name: chat-ui\n secrets:\n hf_token: ${{ secrets.HF_DOC_BUILD_PUSH }}\n comment_bot_token: ${{ secrets.COMMENT_BOT_TOKEN }}\n" }, { "path": ".gitignore", "content": ".DS_Store\nnode_modules\n/build\n/.svelte-kit\n/package\n.env\n.env.*\nvite.config.js.timestamp-*\nvite.config.ts.timestamp-*\nSECRET_CONFIG\n.idea\n!.env.ci\n!.env\ngcp-*.json\ndb\nmodels/*\n!models/add-your-models-here.txt\n.claude/*\n!.claude/skills/" }, { "path": ".husky/lint-stage-config.js", "content": "export default {\n\t\"*.{js,jsx,ts,tsx}\": [\"prettier --write\", \"eslint --fix\", \"eslint\"],\n\t\"*.json\": [\"prettier --write\"],\n};\n" }, { "path": ".husky/pre-commit", "content": "set -e\nnpx lint-staged --config ./.husky/lint-stage-config.js\n" }, { "path": ".npmrc", "content": "engine-strict=true\n" }, { "path": ".prettierignore", "content": ".DS_Store\nnode_modules\n/build\n/.svelte-kit\n/package\n/chart\n.env\n.env.*\n!.env.example\n\n# Ignore files for PNPM, NPM and YARN\npnpm-lock.yaml\npackage-lock.json\nyarn.lock\n" }, { "path": ".prettierrc", "content": "{\n\t\"useTabs\": true,\n\t\"trailingComma\": \"es5\",\n\t\"printWidth\": 100,\n\t\"plugins\": [\"prettier-plugin-svelte\", \"prettier-plugin-tailwindcss\"],\n\t\"overrides\": [{ \"files\": \"*.svelte\", \"options\": { \"parser\": \"svelte\" } }]\n}\n" }, { "path": ".vscode/launch.json", "content": "{\n\t\"version\": \"0.2.0\",\n\t\"configurations\": [\n\t\t{\n\t\t\t\"command\": \"npm run dev\",\n\t\t\t\"name\": \"Run development server\",\n\t\t\t\"request\": \"launch\",\n\t\t\t\"type\": \"node-terminal\"\n\t\t}\n\t]\n}\n" }, { "path": ".vscode/settings.json", "content": "{\n\t\"editor.formatOnSave\": true,\n\t\"editor.defaultFormatter\": \"esbenp.prettier-vscode\",\n\t\"editor.codeActionsOnSave\": {\n\t\t\"source.fixAll\": \"explicit\"\n\t},\n\t\"eslint.validate\": [\"javascript\", \"svelte\"],\n\t\"[svelte]\": {\n\t\t\"editor.defaultFormatter\": \"esbenp.prettier-vscode\"\n\t},\n\t\"[typescript]\": {\n\t\t\"editor.defaultFormatter\": \"esbenp.prettier-vscode\"\n\t}\n}\n" }, { "path": "CLAUDE.md", "content": "# CLAUDE.md\n\nThis file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.\n\n## Overview\n\nChat UI is a SvelteKit application that provides a chat interface for LLMs. It powers HuggingChat (hf.co/chat). The app speaks exclusively to OpenAI-compatible APIs via `OPENAI_BASE_URL`.\n\n## Commands\n\n```bash\nnpm run dev # Start dev server on localhost:5173\nnpm run build # Production build\nnpm run preview # Preview production build\nnpm run check # TypeScript validation (svelte-kit sync + svelte-check)\nnpm run lint # Check formatting (Prettier) and linting (ESLint)\nnpm run format # Auto-format with Prettier\nnpm run test # Run all tests (Vitest)\n```\n\n### Running a Single Test\n\n```bash\nnpx vitest run path/to/file.spec.ts # Run specific test file\nnpx vitest run -t \"test name\" # Run test by name\nnpx vitest --watch path/to/file.spec.ts # Watch mode for single file\n```\n\n### Test Environments\n\nTests are split into three workspaces (configured in vite.config.ts):\n\n- **Client tests** (`*.svelte.test.ts`): Browser environment with Playwright\n- **SSR tests** (`*.ssr.test.ts`): Node environment for server-side rendering\n- **Server tests** (`*.test.ts`, `*.spec.ts`): Node environment for utilities\n\n## Architecture\n\n### Stack\n\n- **SvelteKit 2** with Svelte 5 (uses runes: `$state`, `$effect`, `$bindable`)\n- **MongoDB** for persistence (auto-fallback to in-memory with MongoMemoryServer when `MONGODB_URL` not set)\n- **TailwindCSS** for styling\n\n### Key Directories\n\n```\nsrc/\n├── lib/\n│ ├── components/ # Svelte components (chat/, mcp/, voice/, icons/)\n│ ├── server/\n│ │ ├── api/utils/ # Shared API helpers (auth, superjson, model/conversation resolvers)\n│ │ ├── textGeneration/ # LLM streaming pipeline\n│ │ ├── mcp/ # Model Context Protocol integration\n│ │ ├── router/ # Smart model routing (Omni)\n│ │ ├── database.ts # MongoDB collections\n│ │ ├── models.ts # Model registry from OPENAI_BASE_URL/models\n│ │ └── auth.ts # OpenID Connect authentication\n│ ├── types/ # TypeScript interfaces (Conversation, Message, User, Model, etc.)\n│ ├── stores/ # Svelte stores for reactive state\n│ └── utils/ # Helpers (tree/, marked.ts, auth.ts, etc.)\n├── routes/ # SvelteKit file-based routing\n│ ├── conversation/[id]/ # Chat page + streaming endpoint\n│ ├── settings/ # User settings pages\n│ ├── api/ # Legacy v1 API endpoints (mcp, transcribe, fetch-url)\n│ ├── api/v2/ # REST API endpoints (+server.ts)\n│ └── r/[id]/ # Shared conversation view\n```\n\n### Text Generation Flow\n\n1. User sends message via `POST /conversation/[id]`\n2. Server validates user, fetches conversation history\n3. Builds message tree structure (see `src/lib/utils/tree/`)\n4. Calls LLM endpoint via OpenAI client\n5. Streams response back, stores in MongoDB\n\n### Model Context Protocol (MCP)\n\nMCP servers are configured via `MCP_SERVERS` env var. When enabled, tools are exposed as OpenAI function calls. The router can auto-select tools-capable models when `LLM_ROUTER_ENABLE_TOOLS=true`.\n\n### LLM Router (Omni)\n\nSmart routing via Arch-Router model. Configured with:\n\n- `LLM_ROUTER_ROUTES_PATH`: JSON file defining routes\n- `LLM_ROUTER_ARCH_BASE_URL`: Router endpoint\n- Shortcuts: multimodal routes bypass router if `LLM_ROUTER_ENABLE_MULTIMODAL=true`\n\n### Database Collections\n\n- `conversations` - Chat sessions with nested messages\n- `users` - User accounts (OIDC-backed)\n- `sessions` - Session data\n- `sharedConversations` - Public share links\n- `settings` - User preferences\n\n## Environment Setup\n\nCopy `.env` to `.env.local` and configure:\n\n```env\nOPENAI_BASE_URL=https://router.huggingface.co/v1\nOPENAI_API_KEY=hf_***\n# MONGODB_URL is optional; omit for in-memory DB persisted to ./db\n```\n\nSee `.env` for full list of variables including router config, MCP servers, auth, and feature flags.\n\n## Code Conventions\n\n- TypeScript strict mode enabled\n- ESLint: no `any`, no non-null assertions\n- Prettier: tabs, 100 char width, Tailwind class sorting\n- Server vs client separation via SvelteKit conventions (`+page.server.ts` vs `+page.ts`)\n\n## Feature Development Checklist\n\nWhen building new features, consider:\n\n1. **HuggingChat vs self-hosted**: Wrap HuggingChat-specific features with `publicConfig.isHuggingChat`\n2. **Settings persistence**: Add new fields to `src/lib/types/Settings.ts`, update API endpoint at `src/routes/api/v2/user/settings/+server.ts`\n3. **Rich dropdowns**: Use `bits-ui` (Select, DropdownMenu) instead of native elements when you need icons/images in options\n4. **Scrollbars**: Use `scrollbar-custom` class for styled scrollbars\n5. **Icons**: Custom icons in `$lib/components/icons/`, use Carbon (`~icons/carbon/*`) or Lucide (`~icons/lucide/*`) for standard icons\n6. **Provider avatars**: Use `PROVIDERS_HUB_ORGS` from `@huggingface/inference` for HF provider avatar URLs\n" }, { "path": "Dockerfile", "content": "# syntax=docker/dockerfile:1\nARG INCLUDE_DB=false\n\nFROM node:24-slim AS base\n\n# install dotenv-cli\nRUN npm install -g dotenv-cli\n\n# switch to a user that works for spaces\nRUN userdel -r node\nRUN useradd -m -u 1000 user\nUSER user\n\nENV HOME=/home/user \\\n PATH=/home/user/.local/bin:$PATH\n\nWORKDIR /app\n\n# add a .env.local if the user doesn't bind a volume to it\nRUN touch /app/.env.local\n\nUSER root\nRUN apt-get update\nRUN apt-get install -y libgomp1 libcurl4 curl dnsutils nano\n\n# ensure npm cache dir exists before adjusting ownership\nRUN mkdir -p /home/user/.npm && chown -R 1000:1000 /home/user/.npm\n\nUSER user\n\n\nCOPY --chown=1000 .env /app/.env\nCOPY --chown=1000 entrypoint.sh /app/entrypoint.sh\nCOPY --chown=1000 package.json /app/package.json\nCOPY --chown=1000 package-lock.json /app/package-lock.json\n\nRUN chmod +x /app/entrypoint.sh\n\nFROM node:24 AS builder\n\nWORKDIR /app\n\nCOPY --link --chown=1000 package-lock.json package.json ./\n\nARG APP_BASE=\nARG PUBLIC_APP_COLOR=\nENV BODY_SIZE_LIMIT=15728640\n\nRUN --mount=type=cache,target=/app/.npm \\\n npm set cache /app/.npm && \\\n npm ci\n\nCOPY --link --chown=1000 . .\n\nRUN git config --global --add safe.directory /app && \\\n npm run build\n\n# mongo image\nFROM mongo:7 AS mongo\n\n# image to be used if INCLUDE_DB is false\nFROM base AS local_db_false\n\n# image to be used if INCLUDE_DB is true\nFROM base AS local_db_true\n\n# copy mongo from the other stage\nCOPY --from=mongo /usr/bin/mongo* /usr/bin/\n\nENV MONGODB_URL=mongodb://localhost:27017\nUSER root\nRUN mkdir -p /data/db\nRUN chown -R 1000:1000 /data/db\nUSER user\n# final image\nFROM local_db_${INCLUDE_DB} AS final\n\n# build arg to determine if the database should be included\nARG INCLUDE_DB=false\nENV INCLUDE_DB=${INCLUDE_DB}\n\n# svelte requires APP_BASE at build time so it must be passed as a build arg\nARG APP_BASE=\nARG PUBLIC_APP_COLOR=\nARG PUBLIC_COMMIT_SHA=\nENV PUBLIC_COMMIT_SHA=${PUBLIC_COMMIT_SHA}\nENV BODY_SIZE_LIMIT=15728640\n\n#import the build & dependencies\nCOPY --from=builder --chown=1000 /app/build /app/build\nCOPY --from=builder --chown=1000 /app/node_modules /app/node_modules\n\nCMD [\"/bin/bash\", \"-c\", \"/app/entrypoint.sh\"]\n" }, { "path": "LICENSE", "content": "Copyright 2018- The Hugging Face team. All rights reserved.\n\n Apache License\n Version 2.0, January 2004\n http://www.apache.org/licenses/\n\n TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION\n\n 1. Definitions.\n\n \"License\" shall mean the terms and conditions for use, reproduction,\n and distribution as defined by Sections 1 through 9 of this document.\n\n \"Licensor\" shall mean the copyright owner or entity authorized by\n the copyright owner that is granting the License.\n\n \"Legal Entity\" shall mean the union of the acting entity and all\n other entities that control, are controlled by, or are under common\n control with that entity. For the purposes of this definition,\n \"control\" means (i) the power, direct or indirect, to cause the\n direction or management of such entity, whether by contract or\n otherwise, or (ii) ownership of fifty percent (50%) or more of the\n outstanding shares, or (iii) beneficial ownership of such entity.\n\n \"You\" (or \"Your\") shall mean an individual or Legal Entity\n exercising permissions granted by this License.\n\n \"Source\" form shall mean the preferred form for making modifications,\n including but not limited to software source code, documentation\n source, and configuration files.\n\n \"Object\" form shall mean any form resulting from mechanical\n transformation or translation of a Source form, including but\n not limited to compiled object code, generated documentation,\n and conversions to other media types.\n\n \"Work\" shall mean the work of authorship, whether in Source or\n Object form, made available under the License, as indicated by a\n copyright notice that is included in or attached to the work\n (an example is provided in the Appendix below).\n\n \"Derivative Works\" shall mean any work, whether in Source or Object\n form, that is based on (or derived from) the Work and for which the\n editorial revisions, annotations, elaborations, or other modifications\n represent, as a whole, an original work of authorship. For the purposes\n of this License, Derivative Works shall not include works that remain\n separable from, or merely link (or bind by name) to the interfaces of,\n the Work and Derivative Works thereof.\n\n \"Contribution\" shall mean any work of authorship, including\n the original version of the Work and any modifications or additions\n to that Work or Derivative Works thereof, that is intentionally\n submitted to Licensor for inclusion in the Work by the copyright owner\n or by an individual or Legal Entity authorized to submit on behalf of\n the copyright owner. For the purposes of this definition, \"submitted\"\n means any form of electronic, verbal, or written communication sent\n to the Licensor or its representatives, including but not limited to\n communication on electronic mailing lists, source code control systems,\n and issue tracking systems that are managed by, or on behalf of, the\n Licensor for the purpose of discussing and improving the Work, but\n excluding communication that is conspicuously marked or otherwise\n designated in writing by the copyright owner as \"Not a Contribution.\"\n\n \"Contributor\" shall mean Licensor and any individual or Legal Entity\n on behalf of whom a Contribution has been received by Licensor and\n subsequently incorporated within the Work.\n\n 2. Grant of Copyright License. Subject to the terms and conditions of\n this License, each Contributor hereby grants to You a perpetual,\n worldwide, non-exclusive, no-charge, royalty-free, irrevocable\n copyright license to reproduce, prepare Derivative Works of,\n publicly display, publicly perform, sublicense, and distribute the\n Work and such Derivative Works in Source or Object form.\n\n 3. Grant of Patent License. Subject to the terms and conditions of\n this License, each Contributor hereby grants to You a perpetual,\n worldwide, non-exclusive, no-charge, royalty-free, irrevocable\n (except as stated in this section) patent license to make, have made,\n use, offer to sell, sell, import, and otherwise transfer the Work,\n where such license applies only to those patent claims licensable\n by such Contributor that are necessarily infringed by their\n Contribution(s) alone or by combination of their Contribution(s)\n with the Work to which such Contribution(s) was submitted. If You\n institute patent litigation against any entity (including a\n cross-claim or counterclaim in a lawsuit) alleging that the Work\n or a Contribution incorporated within the Work constitutes direct\n or contributory patent infringement, then any patent licenses\n granted to You under this License for that Work shall terminate\n as of the date such litigation is filed.\n\n 4. Redistribution. You may reproduce and distribute copies of the\n Work or Derivative Works thereof in any medium, with or without\n modifications, and in Source or Object form, provided that You\n meet the following conditions:\n\n (a) You must give any other recipients of the Work or\n Derivative Works a copy of this License; and\n\n (b) You must cause any modified files to carry prominent notices\n stating that You changed the files; and\n\n (c) You must retain, in the Source form of any Derivative Works\n that You distribute, all copyright, patent, trademark, and\n attribution notices from the Source form of the Work,\n excluding those notices that do not pertain to any part of\n the Derivative Works; and\n\n (d) If the Work includes a \"NOTICE\" text file as part of its\n distribution, then any Derivative Works that You distribute must\n include a readable copy of the attribution notices contained\n within such NOTICE file, excluding those notices that do not\n pertain to any part of the Derivative Works, in at least one\n of the following places: within a NOTICE text file distributed\n as part of the Derivative Works; within the Source form or\n documentation, if provided along with the Derivative Works; or,\n within a display generated by the Derivative Works, if and\n wherever such third-party notices normally appear. The contents\n of the NOTICE file are for informational purposes only and\n do not modify the License. You may add Your own attribution\n notices within Derivative Works that You distribute, alongside\n or as an addendum to the NOTICE text from the Work, provided\n that such additional attribution notices cannot be construed\n as modifying the License.\n\n You may add Your own copyright statement to Your modifications and\n may provide additional or different license terms and conditions\n for use, reproduction, or distribution of Your modifications, or\n for any such Derivative Works as a whole, provided Your use,\n reproduction, and distribution of the Work otherwise complies with\n the conditions stated in this License.\n\n 5. Submission of Contributions. Unless You explicitly state otherwise,\n any Contribution intentionally submitted for inclusion in the Work\n by You to the Licensor shall be under the terms and conditions of\n this License, without any additional terms or conditions.\n Notwithstanding the above, nothing herein shall supersede or modify\n the terms of any separate license agreement you may have executed\n with Licensor regarding such Contributions.\n\n 6. Trademarks. This License does not grant permission to use the trade\n names, trademarks, service marks, or product names of the Licensor,\n except as required for reasonable and customary use in describing the\n origin of the Work and reproducing the content of the NOTICE file.\n\n 7. Disclaimer of Warranty. Unless required by applicable law or\n agreed to in writing, Licensor provides the Work (and each\n Contributor provides its Contributions) on an \"AS IS\" BASIS,\n WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or\n implied, including, without limitation, any warranties or conditions\n of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A\n PARTICULAR PURPOSE. You are solely responsible for determining the\n appropriateness of using or redistributing the Work and assume any\n risks associated with Your exercise of permissions under this License.\n\n 8. Limitation of Liability. In no event and under no legal theory,\n whether in tort (including negligence), contract, or otherwise,\n unless required by applicable law (such as deliberate and grossly\n negligent acts) or agreed to in writing, shall any Contributor be\n liable to You for damages, including any direct, indirect, special,\n incidental, or consequential damages of any character arising as a\n result of this License or out of the use or inability to use the\n Work (including but not limited to damages for loss of goodwill,\n work stoppage, computer failure or malfunction, or any and all\n other commercial damages or losses), even if such Contributor\n has been advised of the possibility of such damages.\n\n 9. Accepting Warranty or Additional Liability. While redistributing\n the Work or Derivative Works thereof, You may choose to offer,\n and charge a fee for, acceptance of support, warranty, indemnity,\n or other liability obligations and/or rights consistent with this\n License. However, in accepting such obligations, You may act only\n on Your own behalf and on Your sole responsibility, not on behalf\n of any other Contributor, and only if You agree to indemnify,\n defend, and hold each Contributor harmless for any liability\n incurred by, or claims asserted against, such Contributor by reason\n of your accepting any such warranty or additional liability.\n\n END OF TERMS AND CONDITIONS\n\n APPENDIX: How to apply the Apache License to your work.\n\n To apply the Apache License to your work, attach the following\n boilerplate notice, with the fields enclosed by brackets \"[]\"\n replaced with your own identifying information. (Don't include\n the brackets!) The text should be enclosed in the appropriate\n comment syntax for the file format. We also recommend that a\n file or class name and description of purpose be included on the\n same \"printed page\" as the copyright notice for easier\n identification within third-party archives.\n\n Copyright [yyyy] [name of copyright owner]\n\n Licensed under the Apache License, Version 2.0 (the \"License\");\n you may not use this file except in compliance with the License.\n You may obtain a copy of the License at\n\n http://www.apache.org/licenses/LICENSE-2.0\n\n Unless required by applicable law or agreed to in writing, software\n distributed under the License is distributed on an \"AS IS\" BASIS,\n WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n See the License for the specific language governing permissions and\n limitations under the License." }, { "path": "PRIVACY.md", "content": "## Privacy\n\n> Last updated: Sep 15, 2025\n\nBasics:\n\n- Sign-in: You authenticate with your Hugging Face account.\n- Conversation history: Stored so you can access past chats; you can delete any conversation at any time from the UI.\n\n🗓 Please also consult huggingface.co's main privacy policy at . To exercise any of your legal privacy rights, please send an email to .\n\n## Data handling and processing\n\nHuggingChat uses Hugging Face’s Inference Providers to access models from multiple partners via a single API. Depending on the model and availability, inference runs with the corresponding provider.\n\n- Inference Providers documentation: \n- Security & Compliance: \n\nSecurity and routing facts\n\n- Hugging Face does not store any user data for training purposes.\n- Hugging Face does not store the request body or the response when routing requests through Hugging Face.\n- Logs are kept for debugging purposes for up to 30 days, but no user data or tokens are stored in those logs.\n- Inference Provider routing uses TLS/SSL to encrypt data in transit.\n- The Hugging Face Hub (which Inference Providers is a feature of) is SOC 2 Type 2 certified. See .\n\nExternal providers are responsible for their own security and data handling. Please consult each provider’s respective security and privacy policies via the Inference Providers documentation linked above.\n\n## Technical details\n\n[![chat-ui](https://img.shields.io/github/stars/huggingface/chat-ui)](https://github.com/huggingface/chat-ui)\n\nThe app is completely open source, and further development takes place on the [huggingface/chat-ui](https://github.com/huggingface/chat-ui) GitHub repo. We're always open to contributions!\n\nYou can find the production configuration for HuggingChat [here](https://github.com/huggingface/chat-ui/blob/main/chart/env/prod.yaml).\n\nHuggingChat connects to the OpenAI‑compatible Inference Providers router at `https://router.huggingface.co/v1` to access models across multiple providers. Provider selection may be automatic or fixed depending on the model configuration.\n\nWe welcome any feedback on this app: please participate in the public discussion at \n\n\n" }, { "path": "README.md", "content": "# Chat UI\n\n![Chat UI repository thumbnail](https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/chat-ui/chat-ui-2026.png)\n\nA chat interface for LLMs. It is a SvelteKit app and it powers the [HuggingChat app on hf.co/chat](https://huggingface.co/chat).\n\n0. [Quickstart](#quickstart)\n1. [Database Options](#database-options)\n2. [Launch](#launch)\n3. [Optional Docker Image](#optional-docker-image)\n4. [Extra parameters](#extra-parameters)\n5. [Building](#building)\n\n> [!NOTE]\n> Chat UI only supports OpenAI-compatible APIs via `OPENAI_BASE_URL` and the `/models` endpoint. Provider-specific integrations (legacy `MODELS` env var, GGUF discovery, embeddings, web-search helpers, etc.) are removed, but any service that speaks the OpenAI protocol (llama.cpp server, Ollama, OpenRouter, etc. will work by default).\n\n> [!NOTE]\n> The old version is still available on the [legacy branch](https://github.com/huggingface/chat-ui/tree/legacy)\n\n## Quickstart\n\nChat UI speaks to OpenAI-compatible APIs only. The fastest way to get running is with the Hugging Face Inference Providers router plus your personal Hugging Face access token.\n\n**Step 1 – Create `.env.local`:**\n\n```env\nOPENAI_BASE_URL=https://router.huggingface.co/v1\nOPENAI_API_KEY=hf_************************\n```\n\n`OPENAI_API_KEY` can come from any OpenAI-compatible endpoint you plan to call. Pick the combo that matches your setup and drop the values into `.env.local`:\n\n| Provider | Example `OPENAI_BASE_URL` | Example key env |\n| --------------------------------------------- | ---------------------------------- | ----------------------------------------------------------------------- |\n| Hugging Face Inference Providers router | `https://router.huggingface.co/v1` | `OPENAI_API_KEY=hf_xxx` (or `HF_TOKEN` legacy alias) |\n| llama.cpp server (`llama.cpp --server --api`) | `http://127.0.0.1:8080/v1` | `OPENAI_API_KEY=sk-local-demo` (any string works; llama.cpp ignores it) |\n| Ollama (with OpenAI-compatible bridge) | `http://127.0.0.1:11434/v1` | `OPENAI_API_KEY=ollama` |\n| OpenRouter | `https://openrouter.ai/api/v1` | `OPENAI_API_KEY=sk-or-v1-...` |\n| Poe | `https://api.poe.com/v1` | `OPENAI_API_KEY=pk_...` |\n\nCheck the root [`.env` template](./.env) for the full list of optional variables you can override.\n\n**Step 2 – Install and launch the dev server:**\n\n```bash\ngit clone https://github.com/huggingface/chat-ui\ncd chat-ui\nnpm install\nnpm run dev -- --open\n```\n\nYou now have Chat UI running locally. Open the browser and start chatting.\n\n## Database Options\n\nChat history, users, settings, files, and stats all live in MongoDB. You can point Chat UI at any MongoDB 6/7 deployment.\n\n> [!TIP]\n> For quick local development, you can skip this section. When `MONGODB_URL` is not set, Chat UI falls back to an embedded MongoDB that persists to `./db`.\n\n### MongoDB Atlas (managed)\n\n1. Create a free cluster at [mongodb.com](https://www.mongodb.com/pricing).\n2. Add your IP (or `0.0.0.0/0` for development) to the network access list.\n3. Create a database user and copy the connection string.\n4. Paste that string into `MONGODB_URL` in `.env.local`. Keep the default `MONGODB_DB_NAME=chat-ui` or change it per environment.\n\nAtlas keeps MongoDB off your laptop, which is ideal for teams or cloud deployments.\n\n### Local MongoDB (container)\n\nIf you prefer to run MongoDB in a container:\n\n```bash\ndocker run -d -p 27017:27017 --name mongo-chatui mongo:latest\n```\n\nThen set `MONGODB_URL=mongodb://localhost:27017` in `.env.local`.\n\n## Launch\n\nAfter configuring your environment variables, start Chat UI with:\n\n```bash\nnpm install\nnpm run dev\n```\n\nThe dev server listens on `http://localhost:5173` by default. Use `npm run build` / `npm run preview` for production builds.\n\n## Optional Docker Image\n\nThe `chat-ui-db` image bundles MongoDB inside the container:\n\n```bash\ndocker run \\\n -p 3000:3000 \\\n -e OPENAI_BASE_URL=https://router.huggingface.co/v1 \\\n -e OPENAI_API_KEY=hf_*** \\\n -v chat-ui-data:/data \\\n ghcr.io/huggingface/chat-ui-db:latest\n```\n\nAll environment variables accepted in `.env.local` can be provided as `-e` flags.\n\n## Extra parameters\n\n### Theming\n\nYou can use a few environment variables to customize the look and feel of chat-ui. These are by default:\n\n```env\nPUBLIC_APP_NAME=ChatUI\nPUBLIC_APP_ASSETS=chatui\nPUBLIC_APP_DESCRIPTION=\"Making the community's best AI chat models available to everyone.\"\nPUBLIC_APP_DATA_SHARING=\n```\n\n- `PUBLIC_APP_NAME` The name used as a title throughout the app.\n- `PUBLIC_APP_ASSETS` Is used to find logos & favicons in `static/$PUBLIC_APP_ASSETS`, current options are `chatui` and `huggingchat`.\n- `PUBLIC_APP_DATA_SHARING` Can be set to 1 to add a toggle in the user settings that lets your users opt-in to data sharing with models creator.\n\n### Models\n\nModels are discovered from `${OPENAI_BASE_URL}/models`, and you can optionally override their metadata via the `MODELS` env var (JSON5). Legacy provider‑specific integrations and GGUF discovery are removed. Authorization uses `OPENAI_API_KEY` (preferred). `HF_TOKEN` remains a legacy alias.\n\n### LLM Router (Optional)\n\nChat UI can perform server-side smart routing using [katanemo/Arch-Router-1.5B](https://huggingface.co/katanemo/Arch-Router-1.5B) as the routing model without running a separate router service. The UI exposes a virtual model alias called \"Omni\" (configurable) that, when selected, chooses the best route/model for each message.\n\n- Provide a routes policy JSON via `LLM_ROUTER_ROUTES_PATH`. No sample file ships with this branch, so you must point the variable to a JSON array you create yourself (for example, commit one in your project like `config/routes.chat.json`). Each route entry needs `name`, `description`, `primary_model`, and optional `fallback_models`.\n- Configure the Arch router selection endpoint with `LLM_ROUTER_ARCH_BASE_URL` (OpenAI-compatible `/chat/completions`) and `LLM_ROUTER_ARCH_MODEL` (e.g. `router/omni`). The Arch call reuses `OPENAI_API_KEY` for auth.\n- Map `other` to a concrete route via `LLM_ROUTER_OTHER_ROUTE` (default: `casual_conversation`). If Arch selection fails, calls fall back to `LLM_ROUTER_FALLBACK_MODEL`.\n- Selection timeout can be tuned via `LLM_ROUTER_ARCH_TIMEOUT_MS` (default 10000).\n- Omni alias configuration: `PUBLIC_LLM_ROUTER_ALIAS_ID` (default `omni`), `PUBLIC_LLM_ROUTER_DISPLAY_NAME` (default `Omni`), and optional `PUBLIC_LLM_ROUTER_LOGO_URL`.\n\nWhen you select Omni in the UI, Chat UI will:\n\n- Call the Arch endpoint once (non-streaming) to pick the best route for the last turns.\n- Emit RouterMetadata immediately (route and actual model used) so the UI can display it.\n- Stream from the selected model via your configured `OPENAI_BASE_URL`. On errors, it tries route fallbacks.\n\nTool and multimodal shortcuts:\n\n- Multimodal: If `LLM_ROUTER_ENABLE_MULTIMODAL=true` and the user sends an image, the router bypasses Arch and uses the model specified in `LLM_ROUTER_MULTIMODAL_MODEL`. Route name: `multimodal`.\n- Tools: If `LLM_ROUTER_ENABLE_TOOLS=true` and the user has at least one MCP server enabled, the router bypasses Arch and uses `LLM_ROUTER_TOOLS_MODEL`. If that model is missing or misconfigured, it falls back to Arch routing. Route name: `agentic`.\n\n### MCP Tools (Optional)\n\nChat UI can call tools exposed by Model Context Protocol (MCP) servers and feed results back to the model using OpenAI function calling. You can preconfigure trusted servers via env, let users add their own, and optionally have the Omni router auto‑select a tools‑capable model.\n\nConfigure servers (base list for all users):\n\n```env\n# JSON array of servers: name, url, optional headers\nMCP_SERVERS=[\n {\"name\": \"Web Search (Exa)\", \"url\": \"https://mcp.exa.ai/mcp\"},\n {\"name\": \"Hugging Face MCP Login\", \"url\": \"https://hf.co/mcp?login\"}\n]\n\n# Forward the signed-in user's Hugging Face token to the official HF MCP login endpoint\n# when no Authorization header is set on that server entry.\nMCP_FORWARD_HF_USER_TOKEN=true\n```\n\nEnable router tool path (Omni):\n\n- Set `LLM_ROUTER_ENABLE_TOOLS=true` and choose a tools‑capable target with `LLM_ROUTER_TOOLS_MODEL=`.\n- The target must support OpenAI tools/function calling. Chat UI surfaces a “tools” badge on models that advertise this; you can also force‑enable it per‑model in settings (see below).\n\nUse tools in the UI:\n\n- Open “MCP Servers” from the top‑right menu or from the `+` menu in the chat input to add servers, toggle them on, and run Health Check. The server card lists available tools.\n- When a model calls a tool, the message shows a compact “tool” block with parameters, a progress bar while running, and the result (or error). Results are also provided back to the model for follow‑up.\n\nPer‑model overrides:\n\n- In Settings → Model, you can toggle “Tool calling (functions)” and “Multimodal input” per model. These overrides apply even if the provider metadata doesn’t advertise the capability.\n\n## Building\n\nTo create a production version of your app:\n\n```bash\nnpm run build\n```\n\nYou can preview the production build with `npm run preview`.\n\n> To deploy your app, you may need to install an [adapter](https://kit.svelte.dev/docs/adapters) for your target environment.\n" }, { "path": "chart/Chart.yaml", "content": "apiVersion: v2\nname: chat-ui\nversion: 0.0.1-latest\ntype: application\nicon: https://huggingface.co/front/assets/huggingface_logo-noborder.svg\n" }, { "path": "chart/env/dev.yaml", "content": "image:\n repository: huggingface\n name: chat-ui\n\n#nodeSelector:\n# role-huggingchat: \"true\"\n#\n#tolerations:\n# - key: \"huggingface.co/huggingchat\"\n# operator: \"Equal\"\n# value: \"true\"\n# effect: \"NoSchedule\"\n\nserviceAccount:\n enabled: true\n create: true\n name: huggingchat-ephemeral\n\ningress:\n enabled: false\n\ningressInternal:\n enabled: true\n path: \"/chat\"\n annotations:\n external-dns.alpha.kubernetes.io/hostname: \"*.chat-dev.huggingface.tech\"\n alb.ingress.kubernetes.io/healthcheck-path: \"/chat/healthcheck\"\n alb.ingress.kubernetes.io/listen-ports: \"[{\\\"HTTP\\\": 80}, {\\\"HTTPS\\\": 443}]\"\n alb.ingress.kubernetes.io/group.name: \"chat-dev-internal-public\"\n alb.ingress.kubernetes.io/load-balancer-name: \"chat-dev-internal-public\"\n alb.ingress.kubernetes.io/ssl-redirect: \"443\"\n alb.ingress.kubernetes.io/tags: \"Env=prod,Project=hub,Terraform=true\"\n alb.ingress.kubernetes.io/target-group-attributes: deregistration_delay.timeout_seconds=30\n alb.ingress.kubernetes.io/target-type: \"ip\"\n alb.ingress.kubernetes.io/certificate-arn: \"arn:aws:acm:us-east-1:707930574880:certificate/bc3eb446-1c04-432c-ac6b-946a88d725da\"\n kubernetes.io/ingress.class: \"alb\"\n\nenvVars:\n TEST: \"test\"\n COUPLE_SESSION_WITH_COOKIE_NAME: \"token\"\n OPENID_SCOPES: \"openid profile inference-api read-mcp read-billing\"\n USE_USER_TOKEN: \"true\"\n MCP_FORWARD_HF_USER_TOKEN: \"true\"\n AUTOMATIC_LOGIN: \"false\"\n\n ADDRESS_HEADER: \"X-Forwarded-For\"\n APP_BASE: \"/chat\"\n ALLOW_IFRAME: \"false\"\n COOKIE_SAMESITE: \"lax\"\n COOKIE_SECURE: \"true\"\n EXPOSE_API: \"true\"\n METRICS_ENABLED: \"true\"\n LOG_LEVEL: \"debug\"\n NODE_LOG_STRUCTURED_DATA: \"true\"\n\n OPENAI_BASE_URL: \"https://router.huggingface.co/v1\"\n PUBLIC_APP_ASSETS: \"huggingchat\"\n PUBLIC_APP_NAME: \"HuggingChat\"\n PUBLIC_APP_DESCRIPTION: \"Making the community's best AI chat models available to everyone\"\n PUBLIC_ORIGIN: \"\"\n PUBLIC_PLAUSIBLE_SCRIPT_URL: \"https://plausible.io/js/pa-Io_oigECawqdlgpf5qvHb.js\"\n\n TASK_MODEL: \"Qwen/Qwen3-4B-Instruct-2507\"\n LLM_ROUTER_ARCH_BASE_URL: \"https://router.huggingface.co/v1\"\n LLM_ROUTER_ROUTES_PATH: \"build/client/chat/huggingchat/routes.chat.json\"\n LLM_ROUTER_ARCH_MODEL: \"katanemo/Arch-Router-1.5B\"\n LLM_ROUTER_OTHER_ROUTE: \"casual_conversation\"\n LLM_ROUTER_ARCH_TIMEOUT_MS: \"10000\"\n LLM_ROUTER_ENABLE_MULTIMODAL: \"true\"\n LLM_ROUTER_MULTIMODAL_MODEL: \"Qwen/Qwen3.5-397B-A17B\"\n LLM_ROUTER_ENABLE_TOOLS: \"true\"\n LLM_ROUTER_TOOLS_MODEL: \"moonshotai/Kimi-K2-Instruct-0905\"\n TRANSCRIPTION_MODEL: \"openai/whisper-large-v3-turbo\"\n MCP_SERVERS: >\n [{\"name\": \"Web Search (Exa)\", \"url\": \"https://mcp.exa.ai/mcp?tools=web_search_exa,get_code_context_exa,crawling_exa\"}, {\"name\": \"Hugging Face\", \"url\": \"https://hf.co/mcp?login\"}]\n MCP_TOOL_TIMEOUT_MS: \"120000\"\n PUBLIC_LLM_ROUTER_DISPLAY_NAME: \"Omni\"\n PUBLIC_LLM_ROUTER_LOGO_URL: \"https://cdn-uploads.huggingface.co/production/uploads/5f17f0a0925b9863e28ad517/C5V0v1xZXv6M7FXsdJH9b.png\"\n PUBLIC_LLM_ROUTER_ALIAS_ID: \"omni\"\n MODELS: >\n [\n { \"id\": \"Qwen/Qwen3.5-9B\", \"description\": \"Dense multimodal hybrid with 262K context excelling at reasoning on-device.\" },\n { \"id\": \"CohereLabs/tiny-aya-global\", \"description\": \"Tiny multilingual assistant covering 70+ languages for on-device deployment.\" },\n { \"id\": \"CohereLabs/tiny-aya-earth\", \"description\": \"Regional Aya for African languages with culturally tuned on-device inference.\" },\n { \"id\": \"CohereLabs/tiny-aya-fire\", \"description\": \"Regional Aya for South Asian languages with culturally tuned on-device inference.\" },\n { \"id\": \"CohereLabs/tiny-aya-water\", \"description\": \"Regional Aya for Asia-Pacific and European multilingual on-device tasks.\" },\n { \"id\": \"Qwen/Qwen3.5-122B-A10B\", \"description\": \"Multimodal MoE excelling at agentic tool use with 1M context and 201 languages.\" },\n { \"id\": \"Qwen/Qwen3.5-35B-A3B\", \"description\": \"Compact multimodal MoE with hybrid DeltaNet, 1M context, and 201 languages.\" },\n { \"id\": \"Qwen/Qwen3.5-27B\", \"description\": \"Dense multimodal hybrid with top-tier reasoning density and 1M context.\" },\n { \"id\": \"Qwen/Qwen3.5-397B-A17B\", \"description\": \"Native multimodal MoE with hybrid attention, 1M context, and 201 languages.\", \"parameters\": { \"max_tokens\": 32768 } },\n { \"id\": \"allenai/Olmo-3.1-32B-Think\", \"description\": \"Updated Olmo Think with extended RL for stronger math, code, and instruction following.\" },\n { \"id\": \"MiniMaxAI/MiniMax-M2.5\", \"description\": \"Frontier 230B MoE agent for top-tier coding, tool calling, and fast inference.\" },\n { \"id\": \"zai-org/GLM-5\", \"description\": \"Flagship 745B MoE for agentic reasoning, coding, and creative writing.\" },\n { \"id\": \"Qwen/Qwen3-VL-235B-A22B-Instruct\", \"description\": \"Flagship Qwen3 vision-language MoE for visual agents, documents, and GUI automation.\" },\n { \"id\": \"google/gemma-3n-E4B-it\", \"description\": \"Mobile-first multimodal Gemma handling text, images, video, and audio on-device.\" },\n { \"id\": \"nvidia/NVIDIA-Nemotron-Nano-9B-v2\", \"description\": \"Hybrid Mamba-Transformer with 128K context and controllable reasoning budget.\" },\n { \"id\": \"mistralai/Mistral-7B-Instruct-v0.2\", \"description\": \"Efficient 7B instruction model with 32K context for dialogue and coding.\" },\n { \"id\": \"Qwen/Qwen3-Coder-Next-FP8\", \"description\": \"FP8 Qwen3-Coder-Next for efficient inference with repository-scale coding agents.\" },\n { \"id\": \"arcee-ai/Trinity-Mini\", \"description\": \"Compact US-built MoE for multi-turn agents, tool use, and structured outputs.\" },\n { \"id\": \"Qwen/Qwen3-Coder-Next\", \"description\": \"Ultra-sparse coding MoE for repository-scale agents with 256K context.\" },\n { \"id\": \"moonshotai/Kimi-K2.5\", \"description\": \"Native multimodal agent with agent swarms for parallel tool orchestration.\" },\n { \"id\": \"allenai/Molmo2-8B\", \"description\": \"Open vision-language model excelling at video understanding, pointing, and object tracking.\" },\n { \"id\": \"zai-org/GLM-4.7-Flash\", \"description\": \"Fast GLM-4.7 variant optimized for lower latency coding and agents.\" },\n { \"id\": \"zai-org/GLM-4.7\", \"description\": \"Flagship GLM MoE for coding, reasoning, and agentic tool use.\" },\n { \"id\": \"zai-org/GLM-4.7-FP8\", \"description\": \"FP8 GLM-4.7 for efficient inference with strong coding.\" },\n { \"id\": \"MiniMaxAI/MiniMax-M2.1\", \"description\": \"MoE agent model with multilingual coding and fast outputs.\" },\n { \"id\": \"XiaomiMiMo/MiMo-V2-Flash\", \"description\": \"Fast MoE reasoning model with speculative decoding for agents.\" },\n { \"id\": \"Qwen/Qwen3-VL-32B-Instruct\", \"description\": \"Vision-language Qwen for documents, GUI agents, and visual reasoning.\" },\n { \"id\": \"allenai/Olmo-3.1-32B-Instruct\", \"description\": \"Fully open chat model strong at tool use and dialogue.\" },\n { \"id\": \"zai-org/AutoGLM-Phone-9B-Multilingual\", \"description\": \"Mobile agent for multilingual Android device automation.\" },\n { \"id\": \"utter-project/EuroLLM-22B-Instruct-2512\", \"description\": \"European multilingual model for all EU languages and translation.\" },\n { \"id\": \"dicta-il/DictaLM-3.0-24B-Thinking\", \"description\": \"Hebrew-English reasoning model with explicit thinking traces for bilingual QA and logic.\" },\n { \"id\": \"EssentialAI/rnj-1-instruct\", \"description\": \"8B code and STEM model rivaling larger models on agentic coding, math, and tool use.\" },\n { \"id\": \"MiniMaxAI/MiniMax-M2\", \"description\": \"Compact MoE model tuned for fast coding, agentic workflows, and long-context chat.\" },\n { \"id\": \"PrimeIntellect/INTELLECT-3-FP8\", \"description\": \"FP8 INTELLECT-3 variant for cheaper frontier-level math, code, and general reasoning.\" },\n { \"id\": \"Qwen/Qwen3-VL-30B-A3B-Instruct\", \"description\": \"Flagship Qwen3 vision-language model for high-accuracy image, text, and video reasoning.\" },\n { \"id\": \"Qwen/Qwen3-VL-30B-A3B-Thinking\", \"description\": \"Thinking-mode Qwen3-VL that emits detailed multimodal reasoning traces for difficult problems.\" },\n { \"id\": \"Qwen/Qwen3-VL-8B-Instruct\", \"description\": \"Smaller Qwen3 vision-language assistant for everyday multimodal chat, captioning, and analysis.\" },\n { \"id\": \"aisingapore/Qwen-SEA-LION-v4-32B-IT\", \"description\": \"SEA-LION v4 Qwen optimized for Southeast Asian languages and regional enterprise workloads.\" },\n { \"id\": \"allenai/Olmo-3-32B-Think\", \"description\": \"Fully open 32B thinking model excelling at stepwise math, coding, and research reasoning.\" },\n { \"id\": \"allenai/Olmo-3-7B-Instruct\", \"description\": \"Lightweight Olmo assistant for instruction following, Q&A, and everyday open-source workflows.\" },\n { \"id\": \"allenai/Olmo-3-7B-Think\", \"description\": \"7B Olmo reasoning model delivering transparent multi-step thinking on modest hardware.\" },\n { \"id\": \"deepcogito/cogito-671b-v2.1\", \"description\": \"Frontier-scale 671B MoE focused on deep reasoning, math proofs, and complex coding.\" },\n { \"id\": \"deepcogito/cogito-671b-v2.1-FP8\", \"description\": \"FP8 Cogito v2.1 making 671B-scale reasoning more affordable to serve and experiment with.\" },\n { \"id\": \"deepseek-ai/DeepSeek-V3.2\", \"description\": \"Latest DeepSeek agent model combining strong reasoning, tool-use, and efficient long-context inference.\" },\n { \"id\": \"moonshotai/Kimi-K2-Thinking\", \"description\": \"Reasoning-focused Kimi K2 variant for deep chain-of-thought and large agentic tool flows.\" },\n { \"id\": \"nvidia/NVIDIA-Nemotron-Nano-12B-v2\", \"description\": \"NVIDIA Nano 12B general assistant for coding, chat, and agents with efficient deployment.\" },\n { \"id\": \"ServiceNow-AI/Apriel-1.6-15b-Thinker\", \"description\": \"15B multimodal reasoning model with efficient thinking for enterprise and coding tasks.\" },\n { \"id\": \"openai/gpt-oss-safeguard-20b\", \"description\": \"Safety-focused gpt-oss variant for content classification, policy enforcement, and LLM output filtering.\" },\n { \"id\": \"zai-org/GLM-4.5\", \"description\": \"Flagship GLM agent model unifying advanced reasoning, coding, and tool-using capabilities.\" },\n { \"id\": \"zai-org/GLM-4.5V-FP8\", \"description\": \"FP8 vision-language GLM-4.5V for efficient multilingual visual QA, understanding, and hybrid reasoning.\" }, \n { \"id\": \"deepseek-ai/DeepSeek-V3.2-Exp\", \"description\": \"Experimental V3.2 release focused on faster, lower-cost inference with strong general reasoning and tool use.\" },\n { \"id\": \"zai-org/GLM-4.6\", \"description\": \"Next-gen GLM with very long context and solid multilingual reasoning; good for agents and tools.\" },\n { \"id\": \"Kwaipilot/KAT-Dev\", \"description\": \"Developer-oriented assistant tuned for coding, debugging, and lightweight agent workflows.\" },\n { \"id\": \"Qwen/Qwen2.5-VL-72B-Instruct\", \"description\": \"Flagship multimodal Qwen (text+image) instruction model for high-accuracy visual reasoning and detailed explanations.\" },\n { \"id\": \"deepseek-ai/DeepSeek-V3.1-Terminus\", \"description\": \"Refined V3.1 variant optimized for reliability on long contexts, structured outputs, and tool use.\" },\n { \"id\": \"Qwen/Qwen3-VL-235B-A22B-Thinking\", \"description\": \"Deliberative multimodal Qwen that can produce step-wise visual+text reasoning traces for complex tasks.\" },\n { \"id\": \"zai-org/GLM-4.6-FP8\", \"description\": \"FP8-optimized GLM-4.6 for faster/cheaper deployment with near-parity quality on most tasks.\" },\n { \"id\": \"zai-org/GLM-4.6V\", \"description\": \"106B vision-language model with 128K context and native tool calling for multimodal agents.\", \"parameters\": { \"max_tokens\": 8192 } },\n { \"id\": \"zai-org/GLM-4.6V-Flash\", \"description\": \"9B lightweight vision model for fast local inference with tool calling and UI understanding.\" },\n { \"id\": \"zai-org/GLM-4.6V-FP8\", \"description\": \"FP8-quantized GLM-4.6V for efficient multimodal deployment with native tool use.\" },\n { \"id\": \"Qwen/Qwen3-235B-A22B-Thinking-2507\", \"description\": \"Deliberative text-only 235B Qwen variant for transparent, step-by-step reasoning on hard problems.\" },\n { \"id\": \"Qwen/Qwen3-Next-80B-A3B-Instruct\", \"description\": \"Instruction tuned Qwen for multilingual reasoning, coding, long contexts.\" },\n { \"id\": \"Qwen/Qwen3-Next-80B-A3B-Thinking\", \"description\": \"Thinking mode Qwen that outputs explicit step by step reasoning.\" },\n { \"id\": \"moonshotai/Kimi-K2-Instruct-0905\", \"description\": \"Instruction MoE strong coding and multi step reasoning, long context.\" },\n { \"id\": \"openai/gpt-oss-20b\", \"description\": \"Efficient open model for reasoning and tool use, runs locally.\" },\n { \"id\": \"swiss-ai/Apertus-8B-Instruct-2509\", \"description\": \"Open, multilingual, trained on compliant data transparent global assistant.\" },\n { \"id\": \"openai/gpt-oss-120b\", \"description\": \"High performing open model suitable for large scale applications.\" },\n { \"id\": \"Qwen/Qwen3-Coder-30B-A3B-Instruct\", \"description\": \"Code specialized Qwen long context strong generation and function calling.\" },\n { \"id\": \"meta-llama/Llama-3.1-8B-Instruct\", \"description\": \"Instruction tuned Llama efficient conversational assistant with improved alignment.\" },\n { \"id\": \"Qwen/Qwen2.5-VL-7B-Instruct\", \"description\": \"Vision language Qwen handles images and text for basic multimodal tasks.\" },\n { \"id\": \"Qwen/Qwen3-30B-A3B-Instruct-2507\", \"description\": \"Instruction tuned Qwen reliable general tasks with long context support.\" },\n { \"id\": \"baidu/ERNIE-4.5-VL-28B-A3B-PT\", \"description\": \"Baidu multimodal MoE strong at complex vision language reasoning.\" },\n { \"id\": \"baidu/ERNIE-4.5-0.3B-PT\", \"description\": \"Tiny efficient Baidu model surprisingly long context for lightweight chat.\" },\n { \"id\": \"deepseek-ai/DeepSeek-R1\", \"description\": \"MoE reasoning model excels at math, logic, coding with steps.\" },\n { \"id\": \"baidu/ERNIE-4.5-21B-A3B-PT\", \"description\": \"Efficient Baidu MoE competitive generation with fewer active parameters.\" },\n { \"id\": \"swiss-ai/Apertus-70B-Instruct-2509\", \"description\": \"Open multilingual model trained on open data transparent and capable.\" },\n { \"id\": \"Qwen/Qwen3-4B-Instruct-2507\", \"description\": \"Compact instruction Qwen great for lightweight assistants and apps.\" },\n { \"id\": \"meta-llama/Llama-3.2-3B-Instruct\", \"description\": \"Small efficient Llama for basic conversations and instructions.\" },\n { \"id\": \"Qwen/Qwen3-Coder-480B-A35B-Instruct\", \"description\": \"Huge Qwen coder repository scale understanding and advanced generation.\" },\n { \"id\": \"meta-llama/Meta-Llama-3-8B-Instruct\", \"description\": \"Aligned, efficient Llama dependable open source assistant tasks.\" },\n { \"id\": \"Qwen/Qwen3-4B-Thinking-2507\", \"description\": \"Small Qwen that emits transparent step by step reasoning.\" },\n { \"id\": \"moonshotai/Kimi-K2-Instruct\", \"description\": \"MoE assistant strong coding, reasoning, agentic tasks, long context.\" },\n { \"id\": \"zai-org/GLM-4.5V\", \"description\": \"Vision language MoE state of the art multimodal reasoning.\" },\n { \"id\": \"zai-org/GLM-4.6\", \"description\": \"Hybrid reasoning model top choice for intelligent agent applications.\" },\n { \"id\": \"deepseek-ai/DeepSeek-V3.1\", \"description\": \"Supports direct and thinking style reasoning within one model.\" },\n { \"id\": \"Qwen/Qwen3-8B\", \"description\": \"Efficient Qwen assistant strong multilingual skills and formatting.\" },\n { \"id\": \"Qwen/Qwen3-30B-A3B-Thinking-2507\", \"description\": \"Thinking mode Qwen explicit reasoning for complex interpretable tasks.\" },\n { \"id\": \"google/gemma-3-27b-it\", \"description\": \"Multimodal Gemma long context strong text and image understanding.\" },\n { \"id\": \"zai-org/GLM-4.5-Air\", \"description\": \"Efficient GLM strong reasoning and tool use at lower cost.\" },\n { \"id\": \"HuggingFaceTB/SmolLM3-3B\", \"description\": \"Small multilingual long context model surprisingly strong reasoning.\" },\n { \"id\": \"Qwen/Qwen3-30B-A3B\", \"description\": \"Qwen base model for general use or further fine tuning.\" },\n { \"id\": \"Qwen/Qwen2.5-7B-Instruct\", \"description\": \"Compact instruction model solid for basic conversation and tasks.\" },\n { \"id\": \"Qwen/Qwen3-32B\", \"description\": \"General purpose Qwen strong for complex queries and dialogues.\" },\n { \"id\": \"Qwen/QwQ-32B\", \"description\": \"Preview Qwen showcasing next generation features and alignment.\" },\n { \"id\": \"Qwen/Qwen3-235B-A22B-Instruct-2507\", \"description\": \"Flagship instruction Qwen near state of the art across domains.\" },\n { \"id\": \"meta-llama/Llama-3.3-70B-Instruct\", \"description\": \"Improved Llama alignment and structure powerful complex conversations.\" },\n { \"id\": \"Qwen/Qwen2.5-VL-32B-Instruct\", \"description\": \"Multimodal Qwen advanced visual reasoning for complex image plus text.\" },\n { \"id\": \"deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B\", \"description\": \"Tiny distilled Qwen stepwise math and logic reasoning.\" },\n { \"id\": \"Qwen/Qwen3-235B-A22B\", \"description\": \"Qwen base at flagship scale ideal for custom fine tuning.\" },\n { \"id\": \"meta-llama/Llama-4-Scout-17B-16E-Instruct\", \"description\": \"Processes text and images excels at summarization and cross modal reasoning.\" },\n { \"id\": \"NousResearch/Hermes-4-70B\", \"description\": \"Steerable assistant strong reasoning and creativity highly helpful.\" },\n { \"id\": \"Qwen/Qwen2.5-Coder-32B-Instruct\", \"description\": \"Code model strong generation and tool use bridges sizes.\" },\n { \"id\": \"katanemo/Arch-Router-1.5B\", \"description\": \"Lightweight router model directs queries to specialized backends.\" },\n { \"id\": \"meta-llama/Llama-3.2-1B-Instruct\", \"description\": \"Ultra small Llama handles basic Q and A and instructions.\" },\n { \"id\": \"deepseek-ai/DeepSeek-R1-Distill-Qwen-7B\", \"description\": \"Distilled Qwen excels at stepwise logic in compact footprint.\" },\n { \"id\": \"deepseek-ai/DeepSeek-V3\", \"description\": \"General language model direct answers strong creative and knowledge tasks.\" },\n { \"id\": \"deepseek-ai/DeepSeek-V3-0324\", \"description\": \"Updated V3 better reasoning and coding strong tool use.\" },\n { \"id\": \"CohereLabs/command-a-translate-08-2025\", \"description\": \"Translation focused Command model high quality multilingual translation.\" },\n { \"id\": \"deepseek-ai/DeepSeek-R1-Distill-Qwen-32B\", \"description\": \"Distilled from R1 strong reasoning standout dense model.\" },\n { \"id\": \"baidu/ERNIE-4.5-VL-424B-A47B-Base-PT\", \"description\": \"Multimodal base text image pretraining for cross modal understanding.\" },\n { \"id\": \"meta-llama/Llama-4-Maverick-17B-128E-Instruct\", \"description\": \"MoE multimodal Llama rivals top vision language models.\" },\n { \"id\": \"Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8\", \"description\": \"Quantized giant coder faster lighter retains advanced code generation.\" },\n { \"id\": \"deepseek-ai/DeepSeek-R1-0528-Qwen3-8B\", \"description\": \"Qwen3 variant with R1 reasoning improvements compact and capable.\" },\n { \"id\": \"deepseek-ai/DeepSeek-R1-0528\", \"description\": \"R1 update improved reasoning, fewer hallucinations, adds function calling.\", \"parameters\": { \"max_tokens\": 32000 } },\n { \"id\": \"Qwen/Qwen3-14B\", \"description\": \"Balanced Qwen good performance and efficiency for assistants.\" },\n { \"id\": \"MiniMaxAI/MiniMax-M1-80k\", \"description\": \"Long context MoE very fast excels at long range reasoning and code.\" },\n { \"id\": \"Qwen/Qwen2.5-Coder-7B-Instruct\", \"description\": \"Efficient coding assistant for lightweight programming tasks.\" },\n { \"id\": \"aisingapore/Gemma-SEA-LION-v4-27B-IT\", \"description\": \"Gemma SEA LION optimized for Southeast Asian languages or enterprise.\" },\n { \"id\": \"CohereLabs/aya-expanse-8b\", \"description\": \"Small Aya Expanse broad knowledge and efficient general reasoning.\" },\n { \"id\": \"baichuan-inc/Baichuan-M2-32B\", \"description\": \"Medical reasoning specialist fine tuned for clinical QA bilingual.\" },\n { \"id\": \"Qwen/Qwen2.5-VL-72B-Instruct\", \"description\": \"Vision language Qwen detailed image interpretation and instructions.\" },\n { \"id\": \"meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8\", \"description\": \"FP8 Maverick efficient deployment retains top multimodal capability.\" },\n { \"id\": \"zai-org/GLM-4.1V-9B-Thinking\", \"description\": \"Vision language with explicit reasoning strong for its size.\" },\n { \"id\": \"zai-org/GLM-4.5-Air-FP8\", \"description\": \"FP8 efficient GLM Air hybrid reasoning with minimal compute.\" },\n { \"id\": \"google/gemma-2-2b-it\", \"description\": \"Small Gemma instruction tuned safe responsible outputs easy deployment.\" },\n { \"id\": \"arcee-ai/AFM-4.5B\", \"description\": \"Enterprise focused model strong CPU performance compliant and practical.\" },\n { \"id\": \"deepseek-ai/DeepSeek-R1-Distill-Llama-8B\", \"description\": \"Llama distilled from R1 strong reasoning and structured outputs.\" },\n { \"id\": \"CohereLabs/aya-vision-8b\", \"description\": \"Vision capable Aya handles images and text for basic multimodal.\" },\n { \"id\": \"NousResearch/Hermes-3-Llama-3.1-405B\", \"description\": \"Highly aligned assistant excels at math, code, QA.\" },\n { \"id\": \"Qwen/Qwen2.5-72B-Instruct\", \"description\": \"Accurate detailed instruction model supports tools and long contexts.\" },\n { \"id\": \"meta-llama/Llama-Guard-4-12B\", \"description\": \"Safety guardrail model filters and enforces content policies.\" },\n { \"id\": \"CohereLabs/command-a-vision-07-2025\", \"description\": \"Command model with image input captioning and visual QA.\" },\n { \"id\": \"nvidia/Llama-3_1-Nemotron-Ultra-253B-v1\", \"description\": \"NVIDIA tuned Llama optimized throughput for research and production.\" },\n { \"id\": \"meta-llama/Meta-Llama-3-70B-Instruct\", \"description\": \"Instruction tuned Llama improved reasoning and reliability over predecessors.\" },\n { \"id\": \"NousResearch/Hermes-4-405B\", \"description\": \"Frontier Hermes hybrid reasoning excels at math, code, creativity.\" },\n { \"id\": \"NousResearch/Hermes-2-Pro-Llama-3-8B\", \"description\": \"Small Hermes highly steerable maximized helpfulness for basics.\" },\n { \"id\": \"google/gemma-2-9b-it\", \"description\": \"Gemma with improved accuracy and context safe, easy to deploy.\" },\n { \"id\": \"Sao10K/L3-8B-Stheno-v3.2\", \"description\": \"Community Llama variant themed tuning and unique conversational style.\" },\n { \"id\": \"deepcogito/cogito-v2-preview-llama-109B-MoE\", \"description\": \"MoE preview advanced reasoning tests DeepCogito v2 fine tuning.\" },\n { \"id\": \"CohereLabs/c4ai-command-r-08-2024\", \"description\": \"Cohere Command variant instruction following with specialized tuning.\" },\n { \"id\": \"baidu/ERNIE-4.5-300B-A47B-Base-PT\", \"description\": \"Large base model foundation for specialized language systems.\" },\n { \"id\": \"CohereLabs/aya-expanse-32b\", \"description\": \"Aya Expanse large comprehensive knowledge and reasoning capabilities.\" },\n { \"id\": \"CohereLabs/c4ai-command-a-03-2025\", \"description\": \"Updated Command assistant improved accuracy and general usefulness.\" },\n { \"id\": \"CohereLabs/command-a-reasoning-08-2025\", \"description\": \"Command variant optimized for complex multi step logical reasoning.\" },\n { \"id\": \"alpindale/WizardLM-2-8x22B\", \"description\": \"Multi expert WizardLM MoE approach for efficient high quality generation.\" },\n { \"id\": \"tokyotech-llm/Llama-3.3-Swallow-70B-Instruct-v0.4\", \"description\": \"Academic fine tune potential multilingual and domain improvements.\" },\n { \"id\": \"deepseek-ai/DeepSeek-R1-Distill-Llama-70B\", \"description\": \"Llama distilled from R1 improved reasoning enterprise friendly.\" },\n { \"id\": \"CohereLabs/c4ai-command-r7b-12-2024\", \"description\": \"Small Command variant research or regional adaptation focus.\" },\n { \"id\": \"Sao10K/L3-70B-Euryale-v2.1\", \"description\": \"Creative community instruct model with distinctive persona.\" },\n { \"id\": \"CohereLabs/aya-vision-32b\", \"description\": \"Larger Aya Vision advanced vision language with detailed reasoning.\" },\n { \"id\": \"meta-llama/Llama-3.1-405B-Instruct\", \"description\": \"Massive instruction model very long context excels at complex tasks.\" },\n { \"id\": \"CohereLabs/c4ai-command-r7b-arabic-02-2025\", \"description\": \"Command tuned for Arabic fluent and culturally appropriate outputs.\" },\n { \"id\": \"Sao10K/L3-8B-Lunaris-v1\", \"description\": \"Community Llama creative role play oriented themed persona.\" },\n { \"id\": \"Qwen/Qwen2.5-Coder-7B\", \"description\": \"Small Qwen coder basic programming assistance for low resource environments.\" },\n { \"id\": \"Qwen/QwQ-32B-Preview\", \"description\": \"Preview Qwen experimental features and architecture refinements.\" },\n { \"id\": \"deepseek-ai/DeepSeek-R1-Distill-Qwen-14B\", \"description\": \"Distilled Qwen mid size strong reasoning and clear steps.\" },\n { \"id\": \"meta-llama/Llama-3.1-70B-Instruct\", \"description\": \"Instruction tuned Llama improved reasoning and factual reliability.\" },\n { \"id\": \"Qwen/Qwen3-235B-A22B-FP8\", \"description\": \"FP8 quantized Qwen flagship efficient access to ultra large capabilities.\" },\n { \"id\": \"zai-org/GLM-4-32B-0414\", \"description\": \"Open licensed GLM matches larger proprietary models on benchmarks.\" },\n { \"id\": \"SentientAGI/Dobby-Unhinged-Llama-3.3-70B\", \"description\": \"Unfiltered candid creative outputs intentionally less restricted behavior.\" },\n { \"id\": \"marin-community/marin-8b-instruct\", \"description\": \"Community tuned assistant helpful conversational everyday tasks.\" },\n { \"id\": \"deepseek-ai/DeepSeek-Prover-V2-671B\", \"description\": \"Specialist for mathematical proofs and formal reasoning workflows.\" },\n { \"id\": \"NousResearch/Hermes-3-Llama-3.1-70B\", \"description\": \"Highly aligned assistant strong complex instruction following.\" },\n { \"id\": \"Qwen/Qwen2.5-Coder-3B-Instruct\", \"description\": \"Tiny coding assistant basic code completions and explanations.\" },\n { \"id\": \"deepcogito/cogito-v2-preview-llama-70B\", \"description\": \"Preview fine tune enhanced reasoning and tool use indications.\" },\n { \"id\": \"deepcogito/cogito-v2-preview-llama-405B\", \"description\": \"Preview at frontier scale tests advanced fine tuning methods.\" },\n { \"id\": \"deepcogito/cogito-v2-preview-deepseek-671B-MoE\", \"description\": \"Experimental blend of DeepCogito and DeepSeek approaches for reasoning.\" }\n ]\n\ninfisical:\n enabled: true\n env: \"ephemeral-us-east-1\"\n\nreplicas: 1\nautoscaling:\n enabled: false\n\nresources:\n requests:\n cpu: 2\n memory: 4Gi\n limits:\n cpu: 4\n memory: 8Gi\n" }, { "path": "chart/env/prod.yaml", "content": "image:\n repository: huggingface\n name: chat-ui\n\nnodeSelector:\n role-huggingchat: \"true\"\n\ntolerations:\n - key: \"huggingface.co/huggingchat\"\n operator: \"Equal\"\n value: \"true\"\n effect: \"NoSchedule\"\n\nserviceAccount:\n enabled: true\n create: true\n name: huggingchat-prod\n\ningress:\n path: \"/chat\"\n annotations:\n alb.ingress.kubernetes.io/healthcheck-path: \"/chat/healthcheck\"\n alb.ingress.kubernetes.io/listen-ports: \"[{\\\"HTTP\\\": 80}, {\\\"HTTPS\\\": 443}]\"\n alb.ingress.kubernetes.io/load-balancer-name: \"hub-utils-prod-cloudfront\"\n alb.ingress.kubernetes.io/group.name: \"hub-utils-prod-cloudfront\"\n alb.ingress.kubernetes.io/scheme: \"internal\"\n alb.ingress.kubernetes.io/ssl-redirect: \"443\"\n alb.ingress.kubernetes.io/tags: \"Env=prod,Project=hub,Terraform=true\"\n alb.ingress.kubernetes.io/target-group-attributes: deregistration_delay.timeout_seconds=30\n alb.ingress.kubernetes.io/target-type: \"ip\"\n alb.ingress.kubernetes.io/certificate-arn: \"arn:aws:acm:us-east-1:707930574880:certificate/5b25b145-75db-4837-b9f3-7f238ba8a9c7,arn:aws:acm:us-east-1:707930574880:certificate/bfdf509c-f44b-400f-b9e1-6f7a861abe91\"\n kubernetes.io/ingress.class: \"alb\"\n\ningressInternal:\n enabled: true\n path: \"/chat\"\n annotations:\n alb.ingress.kubernetes.io/healthcheck-path: \"/chat/healthcheck\"\n alb.ingress.kubernetes.io/listen-ports: \"[{\\\"HTTP\\\": 80}, {\\\"HTTPS\\\": 443}]\"\n alb.ingress.kubernetes.io/group.name: \"hub-prod-internal-public\"\n alb.ingress.kubernetes.io/load-balancer-name: \"hub-prod-internal-public\"\n alb.ingress.kubernetes.io/ssl-redirect: \"443\"\n alb.ingress.kubernetes.io/tags: \"Env=prod,Project=hub,Terraform=true\"\n alb.ingress.kubernetes.io/target-group-attributes: deregistration_delay.timeout_seconds=30\n alb.ingress.kubernetes.io/target-type: \"ip\"\n alb.ingress.kubernetes.io/certificate-arn: \"arn:aws:acm:us-east-1:707930574880:certificate/5b25b145-75db-4837-b9f3-7f238ba8a9c7,arn:aws:acm:us-east-1:707930574880:certificate/bfdf509c-f44b-400f-b9e1-6f7a861abe91\"\n kubernetes.io/ingress.class: \"alb\"\n\nenvVars:\n COUPLE_SESSION_WITH_COOKIE_NAME: \"token\"\n OPENID_SCOPES: \"openid profile inference-api read-mcp read-billing\"\n USE_USER_TOKEN: \"true\"\n MCP_FORWARD_HF_USER_TOKEN: \"true\"\n AUTOMATIC_LOGIN: \"false\"\n\n ADDRESS_HEADER: \"X-Forwarded-For\"\n APP_BASE: \"/chat\"\n ALLOW_IFRAME: \"false\"\n COOKIE_SAMESITE: \"lax\"\n COOKIE_SECURE: \"true\"\n EXPOSE_API: \"true\"\n METRICS_ENABLED: \"true\"\n LOG_LEVEL: \"debug\"\n NODE_LOG_STRUCTURED_DATA: \"true\"\n\n OPENAI_BASE_URL: \"https://router.huggingface.co/v1\"\n PUBLIC_APP_ASSETS: \"huggingchat\"\n PUBLIC_APP_NAME: \"HuggingChat\"\n PUBLIC_APP_DESCRIPTION: \"Making the community's best AI chat models available to everyone\"\n PUBLIC_ORIGIN: \"https://huggingface.co\"\n PUBLIC_PLAUSIBLE_SCRIPT_URL: \"https://plausible.io/js/pa-Io_oigECawqdlgpf5qvHb.js\"\n\n TASK_MODEL: \"Qwen/Qwen3-4B-Instruct-2507\"\n LLM_ROUTER_ARCH_BASE_URL: \"https://router.huggingface.co/v1\"\n LLM_ROUTER_ROUTES_PATH: \"build/client/chat/huggingchat/routes.chat.json\"\n LLM_ROUTER_ARCH_MODEL: \"katanemo/Arch-Router-1.5B\"\n LLM_ROUTER_OTHER_ROUTE: \"casual_conversation\"\n LLM_ROUTER_ARCH_TIMEOUT_MS: \"10000\"\n LLM_ROUTER_ENABLE_MULTIMODAL: \"true\"\n LLM_ROUTER_MULTIMODAL_MODEL: \"Qwen/Qwen3.5-397B-A17B\"\n LLM_ROUTER_ENABLE_TOOLS: \"true\"\n LLM_ROUTER_TOOLS_MODEL: \"moonshotai/Kimi-K2-Instruct-0905\"\n TRANSCRIPTION_MODEL: \"openai/whisper-large-v3-turbo\"\n MCP_SERVERS: >\n [{\"name\": \"Web Search (Exa)\", \"url\": \"https://mcp.exa.ai/mcp?tools=web_search_exa,get_code_context_exa,crawling_exa\"}, {\"name\": \"Hugging Face\", \"url\": \"https://hf.co/mcp?login\"}]\n MCP_TOOL_TIMEOUT_MS: \"120000\"\n PUBLIC_LLM_ROUTER_DISPLAY_NAME: \"Omni\"\n PUBLIC_LLM_ROUTER_LOGO_URL: \"https://cdn-uploads.huggingface.co/production/uploads/5f17f0a0925b9863e28ad517/C5V0v1xZXv6M7FXsdJH9b.png\"\n PUBLIC_LLM_ROUTER_ALIAS_ID: \"omni\"\n MODELS: >\n [\n { \"id\": \"Qwen/Qwen3.5-9B\", \"description\": \"Dense multimodal hybrid with 262K context excelling at reasoning on-device.\" },\n { \"id\": \"CohereLabs/tiny-aya-global\", \"description\": \"Tiny multilingual assistant covering 70+ languages for on-device deployment.\" },\n { \"id\": \"CohereLabs/tiny-aya-earth\", \"description\": \"Regional Aya for African languages with culturally tuned on-device inference.\" },\n { \"id\": \"CohereLabs/tiny-aya-fire\", \"description\": \"Regional Aya for South Asian languages with culturally tuned on-device inference.\" },\n { \"id\": \"CohereLabs/tiny-aya-water\", \"description\": \"Regional Aya for Asia-Pacific and European multilingual on-device tasks.\" },\n { \"id\": \"Qwen/Qwen3.5-122B-A10B\", \"description\": \"Multimodal MoE excelling at agentic tool use with 1M context and 201 languages.\" },\n { \"id\": \"Qwen/Qwen3.5-35B-A3B\", \"description\": \"Compact multimodal MoE with hybrid DeltaNet, 1M context, and 201 languages.\" },\n { \"id\": \"Qwen/Qwen3.5-27B\", \"description\": \"Dense multimodal hybrid with top-tier reasoning density and 1M context.\" },\n { \"id\": \"Qwen/Qwen3.5-397B-A17B\", \"description\": \"Native multimodal MoE with hybrid attention, 1M context, and 201 languages.\", \"parameters\": { \"max_tokens\": 32768 } },\n { \"id\": \"allenai/Olmo-3.1-32B-Think\", \"description\": \"Updated Olmo Think with extended RL for stronger math, code, and instruction following.\" },\n { \"id\": \"MiniMaxAI/MiniMax-M2.5\", \"description\": \"Frontier 230B MoE agent for top-tier coding, tool calling, and fast inference.\" },\n { \"id\": \"zai-org/GLM-5\", \"description\": \"Flagship 745B MoE for agentic reasoning, coding, and creative writing.\" },\n { \"id\": \"Qwen/Qwen3-VL-235B-A22B-Instruct\", \"description\": \"Flagship Qwen3 vision-language MoE for visual agents, documents, and GUI automation.\" },\n { \"id\": \"google/gemma-3n-E4B-it\", \"description\": \"Mobile-first multimodal Gemma handling text, images, video, and audio on-device.\" },\n { \"id\": \"nvidia/NVIDIA-Nemotron-Nano-9B-v2\", \"description\": \"Hybrid Mamba-Transformer with 128K context and controllable reasoning budget.\" },\n { \"id\": \"mistralai/Mistral-7B-Instruct-v0.2\", \"description\": \"Efficient 7B instruction model with 32K context for dialogue and coding.\" },\n { \"id\": \"Qwen/Qwen3-Coder-Next-FP8\", \"description\": \"FP8 Qwen3-Coder-Next for efficient inference with repository-scale coding agents.\" },\n { \"id\": \"arcee-ai/Trinity-Mini\", \"description\": \"Compact US-built MoE for multi-turn agents, tool use, and structured outputs.\" },\n { \"id\": \"Qwen/Qwen3-Coder-Next\", \"description\": \"Ultra-sparse coding MoE for repository-scale agents with 256K context.\" },\n { \"id\": \"moonshotai/Kimi-K2.5\", \"description\": \"Native multimodal agent with agent swarms for parallel tool orchestration.\" },\n { \"id\": \"allenai/Molmo2-8B\", \"description\": \"Open vision-language model excelling at video understanding, pointing, and object tracking.\" },\n { \"id\": \"zai-org/GLM-4.7-Flash\", \"description\": \"Fast GLM-4.7 variant optimized for lower latency coding and agents.\" },\n { \"id\": \"zai-org/GLM-4.7\", \"description\": \"Flagship GLM MoE for coding, reasoning, and agentic tool use.\" },\n { \"id\": \"zai-org/GLM-4.7-FP8\", \"description\": \"FP8 GLM-4.7 for efficient inference with strong coding.\" },\n { \"id\": \"MiniMaxAI/MiniMax-M2.1\", \"description\": \"MoE agent model with multilingual coding and fast outputs.\" },\n { \"id\": \"XiaomiMiMo/MiMo-V2-Flash\", \"description\": \"Fast MoE reasoning model with speculative decoding for agents.\" },\n { \"id\": \"Qwen/Qwen3-VL-32B-Instruct\", \"description\": \"Vision-language Qwen for documents, GUI agents, and visual reasoning.\" },\n { \"id\": \"allenai/Olmo-3.1-32B-Instruct\", \"description\": \"Fully open chat model strong at tool use and dialogue.\" },\n { \"id\": \"zai-org/AutoGLM-Phone-9B-Multilingual\", \"description\": \"Mobile agent for multilingual Android device automation.\" },\n { \"id\": \"utter-project/EuroLLM-22B-Instruct-2512\", \"description\": \"European multilingual model for all EU languages and translation.\" },\n { \"id\": \"dicta-il/DictaLM-3.0-24B-Thinking\", \"description\": \"Hebrew-English reasoning model with explicit thinking traces for bilingual QA and logic.\" },\n { \"id\": \"EssentialAI/rnj-1-instruct\", \"description\": \"8B code and STEM model rivaling larger models on agentic coding, math, and tool use.\" },\n { \"id\": \"MiniMaxAI/MiniMax-M2\", \"description\": \"Compact MoE model tuned for fast coding, agentic workflows, and long-context chat.\" },\n { \"id\": \"PrimeIntellect/INTELLECT-3-FP8\", \"description\": \"FP8 INTELLECT-3 variant for cheaper frontier-level math, code, and general reasoning.\" },\n { \"id\": \"Qwen/Qwen3-VL-30B-A3B-Instruct\", \"description\": \"Flagship Qwen3 vision-language model for high-accuracy image, text, and video reasoning.\" },\n { \"id\": \"Qwen/Qwen3-VL-30B-A3B-Thinking\", \"description\": \"Thinking-mode Qwen3-VL that emits detailed multimodal reasoning traces for difficult problems.\" },\n { \"id\": \"Qwen/Qwen3-VL-8B-Instruct\", \"description\": \"Smaller Qwen3 vision-language assistant for everyday multimodal chat, captioning, and analysis.\" },\n { \"id\": \"aisingapore/Qwen-SEA-LION-v4-32B-IT\", \"description\": \"SEA-LION v4 Qwen optimized for Southeast Asian languages and regional enterprise workloads.\" },\n { \"id\": \"allenai/Olmo-3-32B-Think\", \"description\": \"Fully open 32B thinking model excelling at stepwise math, coding, and research reasoning.\" },\n { \"id\": \"allenai/Olmo-3-7B-Instruct\", \"description\": \"Lightweight Olmo assistant for instruction following, Q&A, and everyday open-source workflows.\" },\n { \"id\": \"allenai/Olmo-3-7B-Think\", \"description\": \"7B Olmo reasoning model delivering transparent multi-step thinking on modest hardware.\" },\n { \"id\": \"deepcogito/cogito-671b-v2.1\", \"description\": \"Frontier-scale 671B MoE focused on deep reasoning, math proofs, and complex coding.\" },\n { \"id\": \"deepcogito/cogito-671b-v2.1-FP8\", \"description\": \"FP8 Cogito v2.1 making 671B-scale reasoning more affordable to serve and experiment with.\" },\n { \"id\": \"deepseek-ai/DeepSeek-V3.2\", \"description\": \"Latest DeepSeek agent model combining strong reasoning, tool-use, and efficient long-context inference.\" },\n { \"id\": \"moonshotai/Kimi-K2-Thinking\", \"description\": \"Reasoning-focused Kimi K2 variant for deep chain-of-thought and large agentic tool flows.\" },\n { \"id\": \"nvidia/NVIDIA-Nemotron-Nano-12B-v2\", \"description\": \"NVIDIA Nano 12B general assistant for coding, chat, and agents with efficient deployment.\" },\n { \"id\": \"ServiceNow-AI/Apriel-1.6-15b-Thinker\", \"description\": \"15B multimodal reasoning model with efficient thinking for enterprise and coding tasks.\" },\n { \"id\": \"openai/gpt-oss-safeguard-20b\", \"description\": \"Safety-focused gpt-oss variant for content classification, policy enforcement, and LLM output filtering.\" },\n { \"id\": \"zai-org/GLM-4.5\", \"description\": \"Flagship GLM agent model unifying advanced reasoning, coding, and tool-using capabilities.\" },\n { \"id\": \"zai-org/GLM-4.5V-FP8\", \"description\": \"FP8 vision-language GLM-4.5V for efficient multilingual visual QA, understanding, and hybrid reasoning.\" }, \n { \"id\": \"deepseek-ai/DeepSeek-V3.2-Exp\", \"description\": \"Experimental V3.2 release focused on faster, lower-cost inference with strong general reasoning and tool use.\" },\n { \"id\": \"zai-org/GLM-4.6\", \"description\": \"Next-gen GLM with very long context and solid multilingual reasoning; good for agents and tools.\" },\n { \"id\": \"Kwaipilot/KAT-Dev\", \"description\": \"Developer-oriented assistant tuned for coding, debugging, and lightweight agent workflows.\" },\n { \"id\": \"Qwen/Qwen2.5-VL-72B-Instruct\", \"description\": \"Flagship multimodal Qwen (text+image) instruction model for high-accuracy visual reasoning and detailed explanations.\" },\n { \"id\": \"deepseek-ai/DeepSeek-V3.1-Terminus\", \"description\": \"Refined V3.1 variant optimized for reliability on long contexts, structured outputs, and tool use.\" },\n { \"id\": \"Qwen/Qwen3-VL-235B-A22B-Thinking\", \"description\": \"Deliberative multimodal Qwen that can produce step-wise visual+text reasoning traces for complex tasks.\" },\n { \"id\": \"zai-org/GLM-4.6-FP8\", \"description\": \"FP8-optimized GLM-4.6 for faster/cheaper deployment with near-parity quality on most tasks.\" },\n { \"id\": \"zai-org/GLM-4.6V\", \"description\": \"106B vision-language model with 128K context and native tool calling for multimodal agents.\", \"parameters\": { \"max_tokens\": 8192 } },\n { \"id\": \"zai-org/GLM-4.6V-Flash\", \"description\": \"9B lightweight vision model for fast local inference with tool calling and UI understanding.\" },\n { \"id\": \"zai-org/GLM-4.6V-FP8\", \"description\": \"FP8-quantized GLM-4.6V for efficient multimodal deployment with native tool use.\" },\n { \"id\": \"Qwen/Qwen3-235B-A22B-Thinking-2507\", \"description\": \"Deliberative text-only 235B Qwen variant for transparent, step-by-step reasoning on hard problems.\" },\n { \"id\": \"Qwen/Qwen3-Next-80B-A3B-Instruct\", \"description\": \"Instruction tuned Qwen for multilingual reasoning, coding, long contexts.\" },\n { \"id\": \"Qwen/Qwen3-Next-80B-A3B-Thinking\", \"description\": \"Thinking mode Qwen that outputs explicit step by step reasoning.\" },\n { \"id\": \"moonshotai/Kimi-K2-Instruct-0905\", \"description\": \"Instruction MoE strong coding and multi step reasoning, long context.\" },\n { \"id\": \"openai/gpt-oss-20b\", \"description\": \"Efficient open model for reasoning and tool use, runs locally.\" },\n { \"id\": \"swiss-ai/Apertus-8B-Instruct-2509\", \"description\": \"Open, multilingual, trained on compliant data transparent global assistant.\" },\n { \"id\": \"openai/gpt-oss-120b\", \"description\": \"High performing open model suitable for large scale applications.\" },\n { \"id\": \"Qwen/Qwen3-Coder-30B-A3B-Instruct\", \"description\": \"Code specialized Qwen long context strong generation and function calling.\" },\n { \"id\": \"meta-llama/Llama-3.1-8B-Instruct\", \"description\": \"Instruction tuned Llama efficient conversational assistant with improved alignment.\" },\n { \"id\": \"Qwen/Qwen2.5-VL-7B-Instruct\", \"description\": \"Vision language Qwen handles images and text for basic multimodal tasks.\" },\n { \"id\": \"Qwen/Qwen3-30B-A3B-Instruct-2507\", \"description\": \"Instruction tuned Qwen reliable general tasks with long context support.\" },\n { \"id\": \"baidu/ERNIE-4.5-VL-28B-A3B-PT\", \"description\": \"Baidu multimodal MoE strong at complex vision language reasoning.\" },\n { \"id\": \"baidu/ERNIE-4.5-0.3B-PT\", \"description\": \"Tiny efficient Baidu model surprisingly long context for lightweight chat.\" },\n { \"id\": \"deepseek-ai/DeepSeek-R1\", \"description\": \"MoE reasoning model excels at math, logic, coding with steps.\" },\n { \"id\": \"baidu/ERNIE-4.5-21B-A3B-PT\", \"description\": \"Efficient Baidu MoE competitive generation with fewer active parameters.\" },\n { \"id\": \"swiss-ai/Apertus-70B-Instruct-2509\", \"description\": \"Open multilingual model trained on open data transparent and capable.\" },\n { \"id\": \"Qwen/Qwen3-4B-Instruct-2507\", \"description\": \"Compact instruction Qwen great for lightweight assistants and apps.\" },\n { \"id\": \"meta-llama/Llama-3.2-3B-Instruct\", \"description\": \"Small efficient Llama for basic conversations and instructions.\" },\n { \"id\": \"Qwen/Qwen3-Coder-480B-A35B-Instruct\", \"description\": \"Huge Qwen coder repository scale understanding and advanced generation.\" },\n { \"id\": \"meta-llama/Meta-Llama-3-8B-Instruct\", \"description\": \"Aligned, efficient Llama dependable open source assistant tasks.\" },\n { \"id\": \"Qwen/Qwen3-4B-Thinking-2507\", \"description\": \"Small Qwen that emits transparent step by step reasoning.\" },\n { \"id\": \"moonshotai/Kimi-K2-Instruct\", \"description\": \"MoE assistant strong coding, reasoning, agentic tasks, long context.\" },\n { \"id\": \"zai-org/GLM-4.5V\", \"description\": \"Vision language MoE state of the art multimodal reasoning.\" },\n { \"id\": \"zai-org/GLM-4.6\", \"description\": \"Hybrid reasoning model top choice for intelligent agent applications.\" },\n { \"id\": \"deepseek-ai/DeepSeek-V3.1\", \"description\": \"Supports direct and thinking style reasoning within one model.\" },\n { \"id\": \"Qwen/Qwen3-8B\", \"description\": \"Efficient Qwen assistant strong multilingual skills and formatting.\" },\n { \"id\": \"Qwen/Qwen3-30B-A3B-Thinking-2507\", \"description\": \"Thinking mode Qwen explicit reasoning for complex interpretable tasks.\" },\n { \"id\": \"google/gemma-3-27b-it\", \"description\": \"Multimodal Gemma long context strong text and image understanding.\" },\n { \"id\": \"zai-org/GLM-4.5-Air\", \"description\": \"Efficient GLM strong reasoning and tool use at lower cost.\" },\n { \"id\": \"HuggingFaceTB/SmolLM3-3B\", \"description\": \"Small multilingual long context model surprisingly strong reasoning.\" },\n { \"id\": \"Qwen/Qwen3-30B-A3B\", \"description\": \"Qwen base model for general use or further fine tuning.\" },\n { \"id\": \"Qwen/Qwen2.5-7B-Instruct\", \"description\": \"Compact instruction model solid for basic conversation and tasks.\" },\n { \"id\": \"Qwen/Qwen3-32B\", \"description\": \"General purpose Qwen strong for complex queries and dialogues.\" },\n { \"id\": \"Qwen/QwQ-32B\", \"description\": \"Preview Qwen showcasing next generation features and alignment.\" },\n { \"id\": \"Qwen/Qwen3-235B-A22B-Instruct-2507\", \"description\": \"Flagship instruction Qwen near state of the art across domains.\" },\n { \"id\": \"meta-llama/Llama-3.3-70B-Instruct\", \"description\": \"Improved Llama alignment and structure powerful complex conversations.\" },\n { \"id\": \"Qwen/Qwen2.5-VL-32B-Instruct\", \"description\": \"Multimodal Qwen advanced visual reasoning for complex image plus text.\" },\n { \"id\": \"deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B\", \"description\": \"Tiny distilled Qwen stepwise math and logic reasoning.\" },\n { \"id\": \"Qwen/Qwen3-235B-A22B\", \"description\": \"Qwen base at flagship scale ideal for custom fine tuning.\" },\n { \"id\": \"meta-llama/Llama-4-Scout-17B-16E-Instruct\", \"description\": \"Processes text and images excels at summarization and cross modal reasoning.\" },\n { \"id\": \"NousResearch/Hermes-4-70B\", \"description\": \"Steerable assistant strong reasoning and creativity highly helpful.\" },\n { \"id\": \"Qwen/Qwen2.5-Coder-32B-Instruct\", \"description\": \"Code model strong generation and tool use bridges sizes.\" },\n { \"id\": \"katanemo/Arch-Router-1.5B\", \"description\": \"Lightweight router model directs queries to specialized backends.\" },\n { \"id\": \"meta-llama/Llama-3.2-1B-Instruct\", \"description\": \"Ultra small Llama handles basic Q and A and instructions.\" },\n { \"id\": \"deepseek-ai/DeepSeek-R1-Distill-Qwen-7B\", \"description\": \"Distilled Qwen excels at stepwise logic in compact footprint.\" },\n { \"id\": \"deepseek-ai/DeepSeek-V3\", \"description\": \"General language model direct answers strong creative and knowledge tasks.\" },\n { \"id\": \"deepseek-ai/DeepSeek-V3-0324\", \"description\": \"Updated V3 better reasoning and coding strong tool use.\" },\n { \"id\": \"CohereLabs/command-a-translate-08-2025\", \"description\": \"Translation focused Command model high quality multilingual translation.\" },\n { \"id\": \"deepseek-ai/DeepSeek-R1-Distill-Qwen-32B\", \"description\": \"Distilled from R1 strong reasoning standout dense model.\" },\n { \"id\": \"baidu/ERNIE-4.5-VL-424B-A47B-Base-PT\", \"description\": \"Multimodal base text image pretraining for cross modal understanding.\" },\n { \"id\": \"meta-llama/Llama-4-Maverick-17B-128E-Instruct\", \"description\": \"MoE multimodal Llama rivals top vision language models.\" },\n { \"id\": \"Qwen/Qwen3-Coder-480B-A35B-Instruct-FP8\", \"description\": \"Quantized giant coder faster lighter retains advanced code generation.\" },\n { \"id\": \"deepseek-ai/DeepSeek-R1-0528-Qwen3-8B\", \"description\": \"Qwen3 variant with R1 reasoning improvements compact and capable.\" },\n { \"id\": \"deepseek-ai/DeepSeek-R1-0528\", \"description\": \"R1 update improved reasoning, fewer hallucinations, adds function calling.\", \"parameters\": { \"max_tokens\": 32000 } },\n { \"id\": \"Qwen/Qwen3-14B\", \"description\": \"Balanced Qwen good performance and efficiency for assistants.\" },\n { \"id\": \"MiniMaxAI/MiniMax-M1-80k\", \"description\": \"Long context MoE very fast excels at long range reasoning and code.\" },\n { \"id\": \"Qwen/Qwen2.5-Coder-7B-Instruct\", \"description\": \"Efficient coding assistant for lightweight programming tasks.\" },\n { \"id\": \"aisingapore/Gemma-SEA-LION-v4-27B-IT\", \"description\": \"Gemma SEA LION optimized for Southeast Asian languages or enterprise.\" },\n { \"id\": \"CohereLabs/aya-expanse-8b\", \"description\": \"Small Aya Expanse broad knowledge and efficient general reasoning.\" },\n { \"id\": \"baichuan-inc/Baichuan-M2-32B\", \"description\": \"Medical reasoning specialist fine tuned for clinical QA bilingual.\" },\n { \"id\": \"Qwen/Qwen2.5-VL-72B-Instruct\", \"description\": \"Vision language Qwen detailed image interpretation and instructions.\" },\n { \"id\": \"meta-llama/Llama-4-Maverick-17B-128E-Instruct-FP8\", \"description\": \"FP8 Maverick efficient deployment retains top multimodal capability.\" },\n { \"id\": \"zai-org/GLM-4.1V-9B-Thinking\", \"description\": \"Vision language with explicit reasoning strong for its size.\" },\n { \"id\": \"zai-org/GLM-4.5-Air-FP8\", \"description\": \"FP8 efficient GLM Air hybrid reasoning with minimal compute.\" },\n { \"id\": \"google/gemma-2-2b-it\", \"description\": \"Small Gemma instruction tuned safe responsible outputs easy deployment.\" },\n { \"id\": \"arcee-ai/AFM-4.5B\", \"description\": \"Enterprise focused model strong CPU performance compliant and practical.\" },\n { \"id\": \"deepseek-ai/DeepSeek-R1-Distill-Llama-8B\", \"description\": \"Llama distilled from R1 strong reasoning and structured outputs.\" },\n { \"id\": \"CohereLabs/aya-vision-8b\", \"description\": \"Vision capable Aya handles images and text for basic multimodal.\" },\n { \"id\": \"NousResearch/Hermes-3-Llama-3.1-405B\", \"description\": \"Highly aligned assistant excels at math, code, QA.\" },\n { \"id\": \"Qwen/Qwen2.5-72B-Instruct\", \"description\": \"Accurate detailed instruction model supports tools and long contexts.\" },\n { \"id\": \"meta-llama/Llama-Guard-4-12B\", \"description\": \"Safety guardrail model filters and enforces content policies.\" },\n { \"id\": \"CohereLabs/command-a-vision-07-2025\", \"description\": \"Command model with image input captioning and visual QA.\" },\n { \"id\": \"nvidia/Llama-3_1-Nemotron-Ultra-253B-v1\", \"description\": \"NVIDIA tuned Llama optimized throughput for research and production.\" },\n { \"id\": \"meta-llama/Meta-Llama-3-70B-Instruct\", \"description\": \"Instruction tuned Llama improved reasoning and reliability over predecessors.\" },\n { \"id\": \"NousResearch/Hermes-4-405B\", \"description\": \"Frontier Hermes hybrid reasoning excels at math, code, creativity.\" },\n { \"id\": \"NousResearch/Hermes-2-Pro-Llama-3-8B\", \"description\": \"Small Hermes highly steerable maximized helpfulness for basics.\" },\n { \"id\": \"google/gemma-2-9b-it\", \"description\": \"Gemma with improved accuracy and context safe, easy to deploy.\" },\n { \"id\": \"Sao10K/L3-8B-Stheno-v3.2\", \"description\": \"Community Llama variant themed tuning and unique conversational style.\" },\n { \"id\": \"deepcogito/cogito-v2-preview-llama-109B-MoE\", \"description\": \"MoE preview advanced reasoning tests DeepCogito v2 fine tuning.\" },\n { \"id\": \"CohereLabs/c4ai-command-r-08-2024\", \"description\": \"Cohere Command variant instruction following with specialized tuning.\" },\n { \"id\": \"baidu/ERNIE-4.5-300B-A47B-Base-PT\", \"description\": \"Large base model foundation for specialized language systems.\" },\n { \"id\": \"CohereLabs/aya-expanse-32b\", \"description\": \"Aya Expanse large comprehensive knowledge and reasoning capabilities.\" },\n { \"id\": \"CohereLabs/c4ai-command-a-03-2025\", \"description\": \"Updated Command assistant improved accuracy and general usefulness.\" },\n { \"id\": \"CohereLabs/command-a-reasoning-08-2025\", \"description\": \"Command variant optimized for complex multi step logical reasoning.\" },\n { \"id\": \"alpindale/WizardLM-2-8x22B\", \"description\": \"Multi expert WizardLM MoE approach for efficient high quality generation.\" },\n { \"id\": \"tokyotech-llm/Llama-3.3-Swallow-70B-Instruct-v0.4\", \"description\": \"Academic fine tune potential multilingual and domain improvements.\" },\n { \"id\": \"deepseek-ai/DeepSeek-R1-Distill-Llama-70B\", \"description\": \"Llama distilled from R1 improved reasoning enterprise friendly.\" },\n { \"id\": \"CohereLabs/c4ai-command-r7b-12-2024\", \"description\": \"Small Command variant research or regional adaptation focus.\" },\n { \"id\": \"Sao10K/L3-70B-Euryale-v2.1\", \"description\": \"Creative community instruct model with distinctive persona.\" },\n { \"id\": \"CohereLabs/aya-vision-32b\", \"description\": \"Larger Aya Vision advanced vision language with detailed reasoning.\" },\n { \"id\": \"meta-llama/Llama-3.1-405B-Instruct\", \"description\": \"Massive instruction model very long context excels at complex tasks.\" },\n { \"id\": \"CohereLabs/c4ai-command-r7b-arabic-02-2025\", \"description\": \"Command tuned for Arabic fluent and culturally appropriate outputs.\" },\n { \"id\": \"Sao10K/L3-8B-Lunaris-v1\", \"description\": \"Community Llama creative role play oriented themed persona.\" },\n { \"id\": \"Qwen/Qwen2.5-Coder-7B\", \"description\": \"Small Qwen coder basic programming assistance for low resource environments.\" },\n { \"id\": \"Qwen/QwQ-32B-Preview\", \"description\": \"Preview Qwen experimental features and architecture refinements.\" },\n { \"id\": \"deepseek-ai/DeepSeek-R1-Distill-Qwen-14B\", \"description\": \"Distilled Qwen mid size strong reasoning and clear steps.\" },\n { \"id\": \"meta-llama/Llama-3.1-70B-Instruct\", \"description\": \"Instruction tuned Llama improved reasoning and factual reliability.\" },\n { \"id\": \"Qwen/Qwen3-235B-A22B-FP8\", \"description\": \"FP8 quantized Qwen flagship efficient access to ultra large capabilities.\" },\n { \"id\": \"zai-org/GLM-4-32B-0414\", \"description\": \"Open licensed GLM matches larger proprietary models on benchmarks.\" },\n { \"id\": \"SentientAGI/Dobby-Unhinged-Llama-3.3-70B\", \"description\": \"Unfiltered candid creative outputs intentionally less restricted behavior.\" },\n { \"id\": \"marin-community/marin-8b-instruct\", \"description\": \"Community tuned assistant helpful conversational everyday tasks.\" },\n { \"id\": \"deepseek-ai/DeepSeek-Prover-V2-671B\", \"description\": \"Specialist for mathematical proofs and formal reasoning workflows.\" },\n { \"id\": \"NousResearch/Hermes-3-Llama-3.1-70B\", \"description\": \"Highly aligned assistant strong complex instruction following.\" },\n { \"id\": \"Qwen/Qwen2.5-Coder-3B-Instruct\", \"description\": \"Tiny coding assistant basic code completions and explanations.\" },\n { \"id\": \"deepcogito/cogito-v2-preview-llama-70B\", \"description\": \"Preview fine tune enhanced reasoning and tool use indications.\" },\n { \"id\": \"deepcogito/cogito-v2-preview-llama-405B\", \"description\": \"Preview at frontier scale tests advanced fine tuning methods.\" },\n { \"id\": \"deepcogito/cogito-v2-preview-deepseek-671B-MoE\", \"description\": \"Experimental blend of DeepCogito and DeepSeek approaches for reasoning.\" }\n ]\n\ninfisical:\n enabled: true\n env: \"prod-us-east-1\"\n\nautoscaling:\n enabled: true\n minReplicas: 2\n maxReplicas: 30\n targetMemoryUtilizationPercentage: \"50\"\n targetCPUUtilizationPercentage: \"50\"\n\nresources:\n requests:\n cpu: 2\n memory: 4Gi\n limits:\n cpu: 4\n memory: 8Gi\n" }, { "path": "chart/templates/_helpers.tpl", "content": "{{- define \"name\" -}}\n{{- default $.Release.Name | trunc 63 | trimSuffix \"-\" -}}\n{{- end -}}\n\n{{- define \"app.name\" -}}\nchat-ui\n{{- end -}}\n\n{{- define \"labels.standard\" -}}\nrelease: {{ $.Release.Name | quote }}\nheritage: {{ $.Release.Service | quote }}\nchart: \"{{ include \"name\" . }}\"\napp: \"{{ include \"app.name\" . }}\"\n{{- end -}}\n\n{{- define \"labels.resolver\" -}}\nrelease: {{ $.Release.Name | quote }}\nheritage: {{ $.Release.Service | quote }}\nchart: \"{{ include \"name\" . }}\"\napp: \"{{ include \"app.name\" . }}-resolver\"\n{{- end -}}\n\n" }, { "path": "chart/templates/config.yaml", "content": "apiVersion: v1\nkind: ConfigMap\nmetadata:\n labels: {{ include \"labels.standard\" . | nindent 4 }}\n name: {{ include \"name\" . }}\n namespace: {{ .Release.Namespace }}\ndata:\n {{- range $key, $value := $.Values.envVars }}\n {{ $key }}: {{ $value | quote }}\n {{- end }}\n" }, { "path": "chart/templates/deployment.yaml", "content": "apiVersion: apps/v1\nkind: Deployment\nmetadata:\n labels: {{ include \"labels.standard\" . | nindent 4 }}\n name: {{ include \"name\" . }}\n namespace: {{ .Release.Namespace }}\n {{- if .Values.infisical.enabled }}\n annotations:\n secrets.infisical.com/auto-reload: \"true\"\n {{- end }}\nspec:\n progressDeadlineSeconds: 600\n {{- if not $.Values.autoscaling.enabled }}\n replicas: {{ .Values.replicas }}\n {{- end }}\n revisionHistoryLimit: 10\n selector:\n matchLabels: {{ include \"labels.standard\" . | nindent 6 }}\n strategy:\n rollingUpdate:\n maxSurge: 25%\n maxUnavailable: 25%\n type: RollingUpdate\n template:\n metadata:\n labels: {{ include \"labels.standard\" . | nindent 8 }}\n annotations:\n checksum/config: {{ include (print $.Template.BasePath \"/config.yaml\") . | sha256sum }}\n {{- if $.Values.envVars.NODE_LOG_STRUCTURED_DATA }}\n co.elastic.logs/json.expand_keys: \"true\"\n {{- end }}\n spec:\n {{- if .Values.serviceAccount.enabled }}\n serviceAccountName: \"{{ .Values.serviceAccount.name | default (include \"name\" .) }}\"\n {{- end }}\n containers:\n - name: chat-ui\n image: \"{{ .Values.image.repository }}/{{ .Values.image.name }}:{{ .Values.image.tag }}\"\n imagePullPolicy: {{ .Values.image.pullPolicy }}\n readinessProbe:\n failureThreshold: 30\n periodSeconds: 10\n httpGet:\n path: {{ $.Values.envVars.APP_BASE | default \"\" }}/healthcheck\n port: {{ $.Values.envVars.APP_PORT | default 3000 | int }}\n livenessProbe:\n failureThreshold: 30\n periodSeconds: 10\n httpGet:\n path: {{ $.Values.envVars.APP_BASE | default \"\" }}/healthcheck\n port: {{ $.Values.envVars.APP_PORT | default 3000 | int }}\n ports:\n - containerPort: {{ $.Values.envVars.APP_PORT | default 3000 | int }}\n name: http\n protocol: TCP\n {{- if eq \"true\" $.Values.envVars.METRICS_ENABLED }}\n - containerPort: {{ $.Values.envVars.METRICS_PORT | default 5565 | int }}\n name: metrics\n protocol: TCP\n {{- end }}\n resources: {{ toYaml .Values.resources | nindent 12 }}\n {{- with $.Values.extraEnv }}\n env:\n {{- toYaml . | nindent 14 }}\n {{- end }}\n envFrom:\n - configMapRef:\n name: {{ include \"name\" . }}\n {{- if $.Values.infisical.enabled }}\n - secretRef:\n name: {{ include \"name\" $ }}-secs\n {{- end }}\n {{- with $.Values.extraEnvFrom }}\n {{- toYaml . | nindent 14 }}\n {{- end }}\n nodeSelector: {{ toYaml .Values.nodeSelector | nindent 8 }}\n tolerations: {{ toYaml .Values.tolerations | nindent 8 }}\n volumes:\n - name: config\n configMap:\n name: {{ include \"name\" . }}\n" }, { "path": "chart/templates/hpa.yaml", "content": "{{- if $.Values.autoscaling.enabled }}\napiVersion: autoscaling/v2\nkind: HorizontalPodAutoscaler\nmetadata:\n labels: {{ include \"labels.standard\" . | nindent 4 }}\n name: {{ include \"name\" . }}\n namespace: {{ .Release.Namespace }}\nspec:\n scaleTargetRef:\n apiVersion: apps/v1\n kind: Deployment\n name: {{ include \"name\" . }}\n minReplicas: {{ $.Values.autoscaling.minReplicas }}\n maxReplicas: {{ $.Values.autoscaling.maxReplicas }}\n metrics:\n {{- if ne \"\" $.Values.autoscaling.targetMemoryUtilizationPercentage }}\n - type: Resource\n resource:\n name: memory\n target:\n type: Utilization\n averageUtilization: {{ $.Values.autoscaling.targetMemoryUtilizationPercentage | int }}\n {{- end }}\n {{- if ne \"\" $.Values.autoscaling.targetCPUUtilizationPercentage }}\n - type: Resource\n resource:\n name: cpu\n target:\n type: Utilization\n averageUtilization: {{ $.Values.autoscaling.targetCPUUtilizationPercentage | int }}\n {{- end }}\n behavior:\n scaleDown:\n stabilizationWindowSeconds: 600\n policies:\n - type: Percent\n value: 10\n periodSeconds: 60\n scaleUp:\n stabilizationWindowSeconds: 0\n policies:\n - type: Pods\n value: 1\n periodSeconds: 30\n{{- end }}\n" }, { "path": "chart/templates/infisical.yaml", "content": "{{- if .Values.infisical.enabled }}\napiVersion: secrets.infisical.com/v1alpha1\nkind: InfisicalSecret\nmetadata:\n name: {{ include \"name\" $ }}-infisical-secret\n namespace: {{ $.Release.Namespace }}\nspec:\n authentication:\n universalAuth:\n credentialsRef:\n secretName: {{ .Values.infisical.operatorSecretName | quote }}\n secretNamespace: {{ .Values.infisical.operatorSecretNamespace | quote }}\n secretsScope:\n envSlug: {{ .Values.infisical.env | quote }}\n projectSlug: {{ .Values.infisical.project | quote }}\n secretsPath: /\n hostAPI: {{ .Values.infisical.url | quote }}\n managedSecretReference:\n creationPolicy: Owner\n secretName: {{ include \"name\" $ }}-secs\n secretNamespace: {{ .Release.Namespace | quote }}\n secretType: Opaque\n resyncInterval: {{ .Values.infisical.resyncInterval }}\n{{- end }}\n" }, { "path": "chart/templates/ingress-internal.yaml", "content": "{{- if $.Values.ingressInternal.enabled }}\napiVersion: networking.k8s.io/v1\nkind: Ingress\nmetadata:\n annotations: {{ toYaml .Values.ingressInternal.annotations | nindent 4 }}\n labels: {{ include \"labels.standard\" . | nindent 4 }}\n name: {{ include \"name\" . }}-internal\n namespace: {{ .Release.Namespace }}\nspec:\n {{ if $.Values.ingressInternal.className }}\n ingressClassName: {{ .Values.ingressInternal.className }}\n {{ end }}\n {{- with .Values.ingressInternal.tls }}\n tls:\n - hosts:\n - {{ $.Values.domain | quote }}\n {{- with .secretName }}\n secretName: {{ . }}\n {{- end }}\n {{- end }}\n rules:\n - host: {{ .Values.domain }}\n http:\n paths:\n - backend:\n service:\n name: {{ include \"name\" . }}\n port:\n name: http\n path: {{ $.Values.ingressInternal.path | default \"/\" }}\n pathType: Prefix\n{{- end }}\n" }, { "path": "chart/templates/ingress.yaml", "content": "{{- if $.Values.ingress.enabled }}\napiVersion: networking.k8s.io/v1\nkind: Ingress\nmetadata:\n annotations: {{ toYaml .Values.ingress.annotations | nindent 4 }}\n labels: {{ include \"labels.standard\" . | nindent 4 }}\n name: {{ include \"name\" . }}\n namespace: {{ .Release.Namespace }}\nspec:\n {{ if $.Values.ingress.className }}\n ingressClassName: {{ .Values.ingress.className }}\n {{ end }}\n {{- with .Values.ingress.tls }}\n tls:\n - hosts:\n - {{ $.Values.domain | quote }}\n {{- with .secretName }}\n secretName: {{ . }}\n {{- end }}\n {{- end }}\n rules:\n - host: {{ .Values.domain }}\n http:\n paths:\n - backend:\n service:\n name: {{ include \"name\" . }}\n port:\n name: http\n path: {{ $.Values.ingress.path | default \"/\" }}\n pathType: Prefix\n{{- end }}\n" }, { "path": "chart/templates/network-policy.yaml", "content": "{{- if $.Values.networkPolicy.enabled }}\napiVersion: networking.k8s.io/v1\nkind: NetworkPolicy\nmetadata:\n name: {{ include \"name\" . }}\n namespace: {{ .Release.Namespace }}\nspec:\n egress:\n - ports:\n - port: 53\n protocol: UDP\n to:\n - namespaceSelector:\n matchLabels:\n kubernetes.io/metadata.name: kube-system\n podSelector:\n matchLabels:\n k8s-app: kube-dns\n - to:\n {{- range $ip := .Values.networkPolicy.allowedBlocks }}\n - ipBlock:\n cidr: {{ $ip | quote }}\n {{- end }}\n - to:\n - ipBlock:\n cidr: 0.0.0.0/0\n except:\n - 10.0.0.0/8\n - 172.16.0.0/12\n - 192.168.0.0/16\n - 169.254.169.254/32\n podSelector:\n matchLabels: {{ include \"labels.standard\" . | nindent 6 }}\n policyTypes:\n - Egress\n{{- end }}\n" }, { "path": "chart/templates/service-account.yaml", "content": "{{- if and .Values.serviceAccount.enabled .Values.serviceAccount.create }}\napiVersion: v1\nkind: ServiceAccount\nautomountServiceAccountToken: {{ .Values.serviceAccount.automountServiceAccountToken }}\nmetadata:\n name: \"{{ .Values.serviceAccount.name | default (include \"name\" .) }}\"\n namespace: {{ .Release.Namespace }}\n labels: {{ include \"labels.standard\" . | nindent 4 }}\n {{- with .Values.serviceAccount.annotations }}\n annotations:\n {{- toYaml . | nindent 4 }}\n {{- end }}\n{{- end }}\n" }, { "path": "chart/templates/service-monitor.yaml", "content": "{{- if eq \"true\" $.Values.envVars.METRICS_ENABLED }}\napiVersion: monitoring.coreos.com/v1\nkind: ServiceMonitor\nmetadata:\n labels: {{ include \"labels.standard\" . | nindent 4 }}\n name: {{ include \"name\" . }}\n namespace: {{ .Release.Namespace }}\nspec:\n selector:\n matchLabels: {{ include \"labels.standard\" . | nindent 6 }}\n endpoints:\n - port: metrics\n path: /metrics\n interval: 10s\n scheme: http\t\n scrapeTimeout: 10s\n{{- end }}\n" }, { "path": "chart/templates/service.yaml", "content": "apiVersion: v1\nkind: Service\nmetadata:\n name: \"{{ include \"name\" . }}\"\n annotations: {{ toYaml .Values.service.annotations | nindent 4 }}\n namespace: {{ .Release.Namespace }}\n labels: {{ include \"labels.standard\" . | nindent 4 }}\nspec:\n ports:\n - name: http\n port: 80\n protocol: TCP\n targetPort: http\n {{- if eq \"true\" $.Values.envVars.METRICS_ENABLED }}\n - name: metrics\n port: {{ $.Values.envVars.METRICS_PORT | default 5565 | int }}\n protocol: TCP\n targetPort: metrics\n {{- end }}\n selector: {{ include \"labels.standard\" . | nindent 4 }}\n type: {{.Values.service.type}}\n" }, { "path": "chart/values.yaml", "content": "image:\n repository: ghcr.io/huggingface\n name: chat-ui\n tag: 0.0.0-latest\n pullPolicy: IfNotPresent\n\nreplicas: 3\n\ndomain: huggingface.co\n\nnetworkPolicy:\n enabled: false\n allowedBlocks: []\n\nservice:\n type: NodePort\n annotations: { }\n\nserviceAccount:\n enabled: false\n create: false\n name: \"\"\n automountServiceAccountToken: true\n annotations: { }\n\ningress:\n enabled: true\n path: \"/\"\n annotations: { }\n # className: \"nginx\"\n tls: { }\n # secretName: XXX\n\ningressInternal:\n enabled: false\n path: \"/\"\n annotations: { }\n # className: \"nginx\"\n tls: { }\n\nresources:\n requests:\n cpu: 2\n memory: 4Gi\n limits:\n cpu: 2\n memory: 4Gi\nnodeSelector: {}\ntolerations: []\n\nenvVars: { }\n\ninfisical:\n enabled: false\n env: \"\"\n project: \"huggingchat-v2-a1\"\n url: \"\"\n resyncInterval: 60\n operatorSecretName: \"huggingchat-operator-secrets\"\n operatorSecretNamespace: \"hub-utils\"\n\n# Allow to environment injections on top or instead of infisical\nextraEnvFrom: []\nextraEnv: []\n\nautoscaling:\n enabled: false\n minReplicas: 1\n maxReplicas: 2\n targetMemoryUtilizationPercentage: \"\"\n targetCPUUtilizationPercentage: \"\"\n\n## Metrics removed; monitoring configuration no longer used\n" }, { "path": "docker-compose.yml", "content": "# For development only\n# Set MONGODB_URL=mongodb://localhost:27017 in .env.local to use this container\nservices:\n mongo:\n image: mongo:8\n hostname: mongodb\n ports:\n - ${LOCAL_MONGO_PORT:-27017}:27017\n command: --replSet rs0 --bind_ip_all #--setParameter notablescan=1\n mem_limit: \"5g\"\n mem_reservation: \"3g\"\n healthcheck:\n # need to specify the hostname here because the default is the container name, and we run the app outside of docker\n test: test $$(mongosh --quiet --eval 'try {rs.status().ok} catch(e) {rs.initiate({_id:\"rs0\",members:[{_id:0,host:\"127.0.0.1:${LOCAL_MONGO_PORT:-27017}\"}]}).ok}') -eq 1\n interval: 5s\n volumes:\n - mongodb-data:/data/db\n restart: always\n\nvolumes:\n mongodb-data:\n" }, { "path": "docs/source/_toctree.yml", "content": "- local: index\n title: Chat UI\n- title: Installation\n sections:\n - local: installation/local\n title: Local\n - local: installation/docker\n title: Docker\n - local: installation/helm\n title: Helm\n- title: Configuration\n sections:\n - local: configuration/overview\n title: Overview\n - local: configuration/theming\n title: Theming\n - local: configuration/open-id\n title: OpenID\n - local: configuration/mcp-tools\n title: MCP Tools\n - local: configuration/llm-router\n title: LLM Router\n - local: configuration/metrics\n title: Metrics\n - local: configuration/common-issues\n title: Common Issues\n- title: Developing\n sections:\n - local: developing/architecture\n title: Architecture\n" }, { "path": "docs/source/configuration/common-issues.md", "content": "# Common Issues\n\n## 403: You don't have access to this conversation\n\nThis usually happens when running Chat UI over HTTP without proper cookie configuration.\n\n**Recommended:** Set up a reverse proxy (NGINX, Caddy) to handle HTTPS.\n\n**Alternative:** If you must run over HTTP, configure cookies:\n\n```ini\nCOOKIE_SECURE=false\nCOOKIE_SAMESITE=lax\n```\n\nAlso ensure `PUBLIC_ORIGIN` matches your actual URL:\n\n```ini\nPUBLIC_ORIGIN=http://localhost:5173\n```\n\n## Models not loading\n\nIf models aren't appearing in the UI:\n\n1. Verify `OPENAI_BASE_URL` is correct and accessible\n2. Check that `OPENAI_API_KEY` is valid\n3. Ensure the endpoint returns models at `${OPENAI_BASE_URL}/models`\n\n## Database connection errors\n\nFor development, you can skip MongoDB entirely - Chat UI will use an embedded database.\n\nFor production, verify:\n\n- `MONGODB_URL` is a valid connection string\n- Your IP is whitelisted (for MongoDB Atlas)\n- The database user has read/write permissions\n" }, { "path": "docs/source/configuration/llm-router.md", "content": "# LLM Router\n\nChat UI includes an intelligent routing system that automatically selects the best model for each request. When enabled, users see a virtual \"Omni\" model that routes to specialized models based on the conversation context.\n\nThe router uses [katanemo/Arch-Router-1.5B](https://huggingface.co/katanemo/Arch-Router-1.5B) for route selection.\n\n## Configuration\n\n### Basic Setup\n\n```ini\n# Arch router endpoint (OpenAI-compatible)\nLLM_ROUTER_ARCH_BASE_URL=https://router.huggingface.co/v1\nLLM_ROUTER_ARCH_MODEL=katanemo/Arch-Router-1.5B\n\n# Path to your routes policy JSON\nLLM_ROUTER_ROUTES_PATH=./config/routes.json\n```\n\n### Routes Policy\n\nCreate a JSON file defining your routes. Each route specifies:\n\n```json\n[\n\t{\n\t\t\"name\": \"coding\",\n\t\t\"description\": \"Programming, debugging, code review\",\n\t\t\"primary_model\": \"Qwen/Qwen3-Coder-480B-A35B-Instruct\",\n\t\t\"fallback_models\": [\"meta-llama/Llama-3.3-70B-Instruct\"]\n\t},\n\t{\n\t\t\"name\": \"casual_conversation\",\n\t\t\"description\": \"General chat, questions, explanations\",\n\t\t\"primary_model\": \"meta-llama/Llama-3.3-70B-Instruct\"\n\t}\n]\n```\n\n### Fallback Behavior\n\n```ini\n# Route to use when Arch returns \"other\"\nLLM_ROUTER_OTHER_ROUTE=casual_conversation\n\n# Model to use if Arch selection fails entirely\nLLM_ROUTER_FALLBACK_MODEL=meta-llama/Llama-3.3-70B-Instruct\n\n# Selection timeout (milliseconds)\nLLM_ROUTER_ARCH_TIMEOUT_MS=10000\n```\n\n## Multimodal Routing\n\nWhen a user sends an image, the router can bypass Arch and route directly to a vision model:\n\n```ini\nLLM_ROUTER_ENABLE_MULTIMODAL=true\nLLM_ROUTER_MULTIMODAL_MODEL=meta-llama/Llama-3.2-90B-Vision-Instruct\n```\n\n## Tools Routing\n\nWhen a user has MCP servers enabled, the router can automatically select a tools-capable model:\n\n```ini\nLLM_ROUTER_ENABLE_TOOLS=true\nLLM_ROUTER_TOOLS_MODEL=meta-llama/Llama-3.3-70B-Instruct\n```\n\n## UI Customization\n\nCustomize how the router appears in the model selector:\n\n```ini\nPUBLIC_LLM_ROUTER_ALIAS_ID=omni\nPUBLIC_LLM_ROUTER_DISPLAY_NAME=Omni\nPUBLIC_LLM_ROUTER_LOGO_URL=https://example.com/logo.png\n```\n\n## How It Works\n\nWhen a user selects Omni:\n\n1. Chat UI sends the conversation context to the Arch router\n2. Arch analyzes the content and returns a route name\n3. Chat UI maps the route to the corresponding model\n4. The request streams from the selected model\n5. On errors, fallback models are tried in order\n\nThe route selection is displayed in the UI so users can see which model was chosen.\n\n## Message Length Limits\n\nTo optimize router performance, message content is trimmed before sending to Arch:\n\n```ini\n# Max characters for assistant messages (default: 500)\nLLM_ROUTER_MAX_ASSISTANT_LENGTH=500\n\n# Max characters for previous user messages (default: 400)\nLLM_ROUTER_MAX_PREV_USER_LENGTH=400\n```\n\nThe latest user message is never trimmed.\n" }, { "path": "docs/source/configuration/mcp-tools.md", "content": "# MCP Tools\n\nChat UI supports tool calling via the [Model Context Protocol (MCP)](https://modelcontextprotocol.io/). MCP servers expose tools that models can invoke during conversations.\n\n## Server Types\n\nChat UI supports two types of MCP servers:\n\n### Base Servers (Admin-configured)\n\nBase servers are configured by the administrator via environment variables. They appear for all users and can be enabled/disabled per-user but not removed.\n\n```ini\nMCP_SERVERS=[\n {\"name\": \"Web Search (Exa)\", \"url\": \"https://mcp.exa.ai/mcp\"},\n {\"name\": \"Hugging Face\", \"url\": \"https://hf.co/mcp\"}\n]\n```\n\nEach server entry requires:\n\n- `name` - Display name shown in the UI\n- `url` - MCP server endpoint URL\n- `headers` (optional) - Custom headers for authentication\n\n### User Servers (Added from UI)\n\nUsers can add their own MCP servers directly from the UI:\n\n1. Open the chat input and click the **+** button (or go to Settings)\n2. Select **MCP Servers**\n3. Click **Add Server**\n4. Enter the server name and URL\n5. Run **Health Check** to verify connectivity\n\nUser-added servers are stored in the browser and can be removed at any time. They work alongside base servers.\n\n## User Token Forwarding\n\nWhen users are logged in via Hugging Face, you can forward their access token to MCP servers:\n\n```ini\nMCP_FORWARD_HF_USER_TOKEN=true\n```\n\nThis allows MCP servers to access user-specific resources on their behalf.\n\n## Using Tools\n\n1. Enable the servers you want to use from the MCP Servers panel\n2. Start chatting - models will automatically use tools when appropriate\n\n### Model Requirements\n\nNot all models support tool calling. To enable tools for a specific model, add it to your `MODELS` override:\n\n```ini\nMODELS=`[\n {\n \"id\": \"meta-llama/Llama-3.3-70B-Instruct\",\n \"supportsTools\": true\n }\n]`\n```\n\n## Tool Execution Flow\n\nWhen a model decides to use a tool:\n\n1. The model generates a tool call with parameters\n2. Chat UI executes the call against the MCP server\n3. Results are displayed in the chat as a collapsible \"tool\" block\n4. Results are fed back to the model for follow-up responses\n\n## Integration with LLM Router\n\nWhen using the [LLM Router](./llm-router), you can configure automatic routing to a tools-capable model:\n\n```ini\nLLM_ROUTER_ENABLE_TOOLS=true\nLLM_ROUTER_TOOLS_MODEL=meta-llama/Llama-3.3-70B-Instruct\n```\n\nWhen a user has MCP servers enabled and selects the Omni model, the router will automatically use the specified tools model.\n" }, { "path": "docs/source/configuration/metrics.md", "content": "# Metrics\n\nThe server can expose prometheus metrics on port `5565` but is off by default. You may enable the metrics server with `METRICS_ENABLED=true` and change the port with `METRICS_PORT=1234`.\n\n\n\nIn development with `npm run dev`, the metrics server does not shutdown gracefully due to Sveltekit not providing hooks for restart. It's recommended to disable the metrics server in this case.\n\n\n" }, { "path": "docs/source/configuration/open-id.md", "content": "# OpenID\n\nBy default, users are attributed a unique ID based on their browser session. To authenticate users with OpenID Connect, configure the following:\n\n```ini\nOPENID_CLIENT_ID=your_client_id\nOPENID_CLIENT_SECRET=your_client_secret\nOPENID_SCOPES=\"openid profile\"\n```\n\nUse the provider URL for standard OpenID Connect discovery:\n\n```ini\nOPENID_PROVIDER_URL=https://your-provider.com\n```\n\nAdvanced: you can also provide a client metadata document via `OPENID_CONFIG`. This value must be a JSON/JSON5 object (for example, a CIMD document) and is parsed server‑side to populate OpenID settings.\n\n**Redirect URI:** `https://your-domain.com/login/callback`\n\n## Access Control\n\nRestrict access to specific users:\n\n```ini\n# Allow only specific email addresses\nALLOWED_USER_EMAILS=[\"user@example.com\", \"admin@example.com\"]\n\n# Allow all users from specific domains\nALLOWED_USER_DOMAINS=[\"example.com\", \"company.org\"]\n```\n\n## Hugging Face Login\n\nFor Hugging Face authentication, you can use automatic client registration:\n\n```ini\nOPENID_CLIENT_ID=__CIMD__\n```\n\nThis creates an OAuth app automatically when deployed. See the [CIMD spec](https://datatracker.ietf.org/doc/draft-ietf-oauth-client-id-metadata-document/) for details.\n\n## User Token Forwarding\n\nWhen users log in via Hugging Face, you can forward their token for inference:\n\n```ini\nUSE_USER_TOKEN=true\n```\n\n## Auto-Login\n\nForce authentication on all routes:\n\n```ini\nAUTOMATIC_LOGIN=true\n```\n" }, { "path": "docs/source/configuration/overview.md", "content": "# Configuration Overview\n\nChat UI is configured through environment variables. Default values are in `.env`; override them in `.env.local` or via your environment.\n\n## Required Configuration\n\nChat UI connects to any OpenAI-compatible API endpoint:\n\n```ini\nOPENAI_BASE_URL=https://router.huggingface.co/v1\nOPENAI_API_KEY=hf_************************\n```\n\nModels are automatically discovered from `${OPENAI_BASE_URL}/models`. No manual model configuration is required.\n\n## Database\n\n```ini\nMONGODB_URL=mongodb://localhost:27017\nMONGODB_DB_NAME=chat-ui\n```\n\nFor development, `MONGODB_URL` is optional - Chat UI falls back to an embedded MongoDB that persists to `./db`.\n\n## Model Overrides\n\nTo customize model behavior, use the `MODELS` environment variable (JSON5 format):\n\n```ini\nMODELS=`[\n {\n \"id\": \"meta-llama/Llama-3.3-70B-Instruct\",\n \"name\": \"Llama 3.3 70B\",\n \"multimodal\": false,\n \"supportsTools\": true\n }\n]`\n```\n\nOverride properties:\n\n- `id` - Model identifier (must match an ID from the `/models` endpoint)\n- `name` - Display name in the UI\n- `multimodal` - Enable image uploads\n- `supportsTools` - Enable MCP tool calling for models that don’t advertise tool support\n- `parameters` - Override default parameters (temperature, max_tokens, etc.)\n\n## Task Model\n\nSet a specific model for internal tasks (title generation, etc.):\n\n```ini\nTASK_MODEL=meta-llama/Llama-3.1-8B-Instruct\n```\n\nIf not set, the current conversation model is used.\n\n## Voice Transcription\n\nEnable voice input with Whisper:\n\n```ini\nTRANSCRIPTION_MODEL=openai/whisper-large-v3-turbo\nTRANSCRIPTION_BASE_URL=https://router.huggingface.co/hf-inference/models\n```\n\n## Feature Flags\n\n```ini\nLLM_SUMMARIZATION=true # Enable automatic conversation title generation\nENABLE_DATA_EXPORT=true # Allow users to export their data\nALLOW_IFRAME=false # Disallow embedding in iframes (set to true to allow)\n```\n\n## User Authentication\n\nUse OpenID Connect for authentication:\n\n```ini\nOPENID_CLIENT_ID=your_client_id\nOPENID_CLIENT_SECRET=your_client_secret\nOPENID_SCOPES=\"openid profile\"\n```\n\nSee [OpenID configuration](./open-id) for details.\n\n## Environment Variable Reference\n\nSee the [`.env` file](https://github.com/huggingface/chat-ui/blob/main/.env) for the complete list of available options.\n" }, { "path": "docs/source/configuration/theming.md", "content": "# Theming\n\nCustomize the look and feel of Chat UI with these environment variables:\n\n```ini\nPUBLIC_APP_NAME=ChatUI\nPUBLIC_APP_ASSETS=chatui\nPUBLIC_APP_DESCRIPTION=\"Making the community's best AI chat models available to everyone.\"\n```\n\n- `PUBLIC_APP_NAME` - The name used as a title throughout the app\n- `PUBLIC_APP_ASSETS` - Directory for logos & favicons in `static/$PUBLIC_APP_ASSETS`. Options: `chatui`, `huggingchat`\n- `PUBLIC_APP_DESCRIPTION` - Description shown in meta tags and about sections\n\n## Additional Options\n\n```ini\nPUBLIC_APP_DATA_SHARING=1 # Show data sharing opt-in toggle in settings\nPUBLIC_ORIGIN=https://chat.example.com # Your public URL (required for sharing)\n```\n" }, { "path": "docs/source/developing/architecture.md", "content": "# Architecture\n\nThis document provides a high-level overview of the Chat UI codebase. If you're looking to contribute or understand how the codebase works, this is the place for you!\n\n## Overview\n\nChat UI provides a simple interface connecting LLMs to external tools via MCP. The project uses [MongoDB](https://www.mongodb.com/) and [SvelteKit](https://kit.svelte.dev/) with [Tailwind](https://tailwindcss.com/).\n\nKey architectural decisions:\n\n- **OpenAI-compatible only**: All model interactions use the OpenAI API format\n- **MCP for tools**: Tool calling is handled via Model Context Protocol servers\n- **Auto-discovery**: Models are discovered from the `/models` endpoint\n\n## Code Map\n\n### `routes`\n\nAll routes rendered with SSR via SvelteKit. The majority of backend and frontend logic lives here, with shared modules in `lib` (client) and `lib/server` (server).\n\n### `textGeneration`\n\nProvides a standard interface for chat features including model output, tool calls, and streaming. Outputs `MessageUpdate`s for fine-grained status updates (new tokens, tool results, etc.).\n\n### `endpoints`\n\nProvides the streaming interface for OpenAI-compatible endpoints. Models are fetched and cached from `${OPENAI_BASE_URL}/models`.\n\n### `mcp`\n\nImplements MCP client functionality for tool discovery and execution. See [MCP Tools](../configuration/mcp-tools) for configuration.\n\n### `llmRouter`\n\nIntelligent routing logic that selects the best model for each request. Uses the Arch router model for classification. See [LLM Router](../configuration/llm-router) for details.\n\n### `migrations`\n\nMongoDB migrations for maintaining backwards compatibility across schema changes. Any schema changes must include a migration.\n\n## Development\n\n```bash\nnpm install\nnpm run dev\n```\n\nThe dev server runs at `http://localhost:5173` with hot reloading.\n" }, { "path": "docs/source/index.md", "content": "# Chat UI\n\nOpen source chat interface with support for tools, multimodal inputs, and intelligent routing across models. The app uses MongoDB and SvelteKit behind the scenes. Try the live version called [HuggingChat on hf.co/chat](https://huggingface.co/chat) or [setup your own instance](./installation/local).\n\nChat UI connects to any OpenAI-compatible API endpoint, making it work with:\n\n- [Hugging Face Inference Providers](https://huggingface.co/docs/inference-providers)\n- [Ollama](https://ollama.ai)\n- [llama.cpp](https://github.com/ggerganov/llama.cpp)\n- [OpenRouter](https://openrouter.ai)\n- Any other OpenAI-compatible service\n\n**[MCP Tools](./configuration/mcp-tools)**: Function calling via Model Context Protocol (MCP) servers\n\n**[LLM Router](./configuration/llm-router)**: Intelligent routing to select the best model for each request\n\n**[Multimodal](./configuration/overview)**: Image uploads on models that support vision\n\n**[OpenID](./configuration/open-id)**: Optional user authentication via OpenID Connect\n\n## Quickstart\n\n**Step 1 - Create `.env.local`:**\n\n```ini\nOPENAI_BASE_URL=https://router.huggingface.co/v1\nOPENAI_API_KEY=hf_************************\n```\n\nYou can use any OpenAI-compatible endpoint:\n\n| Provider | `OPENAI_BASE_URL` | `OPENAI_API_KEY` |\n| ------------ | ---------------------------------- | ---------------- |\n| Hugging Face | `https://router.huggingface.co/v1` | `hf_xxx` |\n| Ollama | `http://127.0.0.1:11434/v1` | `ollama` |\n| llama.cpp | `http://127.0.0.1:8080/v1` | `sk-local` |\n| OpenRouter | `https://openrouter.ai/api/v1` | `sk-or-v1-xxx` |\n\n**Step 2 - Install and run:**\n\n```bash\ngit clone https://github.com/huggingface/chat-ui\ncd chat-ui\nnpm install\nnpm run dev -- --open\n```\n\nThat's it! Chat UI will automatically discover available models from your endpoint.\n\n> [!TIP]\n> MongoDB is optional for development. When `MONGODB_URL` is not set, Chat UI uses an embedded database that persists to `./db`.\n\nFor production deployments, see the [installation guides](./installation/local).\n" }, { "path": "docs/source/installation/docker.md", "content": "# Running on Docker\n\nPre-built Docker images are available:\n\n- **`ghcr.io/huggingface/chat-ui-db`** - Includes MongoDB (recommended for quick setup)\n- **`ghcr.io/huggingface/chat-ui`** - Requires external MongoDB\n\n## Quick Start (with bundled MongoDB)\n\n```bash\ndocker run -p 3000:3000 \\\n -e OPENAI_BASE_URL=https://router.huggingface.co/v1 \\\n -e OPENAI_API_KEY=hf_*** \\\n -v chat-ui-data:/data \\\n ghcr.io/huggingface/chat-ui-db\n```\n\n## With External MongoDB\n\nIf you have an existing MongoDB instance:\n\n```bash\ndocker run -p 3000:3000 \\\n -e OPENAI_BASE_URL=https://router.huggingface.co/v1 \\\n -e OPENAI_API_KEY=hf_*** \\\n -e MONGODB_URL=mongodb://host.docker.internal:27017 \\\n ghcr.io/huggingface/chat-ui\n```\n\nUse `host.docker.internal` to reach MongoDB running on your host machine, or provide your MongoDB Atlas connection string.\n\n## Using an Environment File\n\nFor more configuration options, use `--env-file` to avoid leaking secrets in shell history:\n\n```bash\ndocker run -p 3000:3000 \\\n --env-file .env.local \\\n -v chat-ui-data:/data \\\n ghcr.io/huggingface/chat-ui-db\n```\n\nSee the [configuration overview](../configuration/overview) for all available environment variables.\n" }, { "path": "docs/source/installation/helm.md", "content": "# Helm\n\n\n\nThe Helm chart is a work in progress and should be considered unstable. Breaking changes may be pushed without migration guides. Contributions welcome!\n\n\n\nFor Kubernetes deployment, use the Helm chart in `/chart`. No chart repository is published, so clone the repository and install by path.\n\n## Installation\n\n```bash\ngit clone https://github.com/huggingface/chat-ui\ncd chat-ui\nhelm install chat-ui ./chart -f values.yaml\n```\n\n## Example values.yaml\n\n```yaml\nreplicas: 1\n\ndomain: example.com\n\nservice:\n type: ClusterIP\n\nresources:\n requests:\n cpu: 100m\n memory: 2Gi\n limits:\n cpu: \"4\"\n memory: 6Gi\n\nenvVars:\n OPENAI_BASE_URL: https://router.huggingface.co/v1\n OPENAI_API_KEY: hf_***\n MONGODB_URL: mongodb://chat-ui-mongo:27017\n```\n\nSee the [configuration overview](../configuration/overview) for all available environment variables.\n" }, { "path": "docs/source/installation/local.md", "content": "# Running Locally\n\n## Quick Start\n\n1. Create a `.env.local` file with your API credentials:\n\n```ini\nOPENAI_BASE_URL=https://router.huggingface.co/v1\nOPENAI_API_KEY=hf_************************\n```\n\n2. Install and run:\n\n```bash\nnpm install\nnpm run dev -- --open\n```\n\nThat's it! Chat UI will discover available models automatically from your endpoint.\n\n## Configuration\n\nChat UI connects to any OpenAI-compatible API. Set `OPENAI_BASE_URL` to your provider:\n\n| Provider | `OPENAI_BASE_URL` |\n| ------------ | ---------------------------------- |\n| Hugging Face | `https://router.huggingface.co/v1` |\n| Ollama | `http://127.0.0.1:11434/v1` |\n| llama.cpp | `http://127.0.0.1:8080/v1` |\n| OpenRouter | `https://openrouter.ai/api/v1` |\n\nSee the [configuration overview](../configuration/overview) for all available options.\n\n## Database\n\nFor **development**, MongoDB is optional. When `MONGODB_URL` is not set, Chat UI uses an embedded MongoDB server that persists data to the `./db` folder.\n\nFor **production**, you should use a dedicated MongoDB instance:\n\n### Option 1: Local MongoDB (Docker)\n\n```bash\ndocker run -d -p 27017:27017 -v mongo-chat-ui:/data --name mongo-chat-ui mongo:latest\n```\n\nThen set `MONGODB_URL=mongodb://localhost:27017` in `.env.local`.\n\n### Option 2: MongoDB Atlas (Managed)\n\nUse [MongoDB Atlas free tier](https://www.mongodb.com/pricing) for a managed database. Copy the connection string to `MONGODB_URL`.\n\n## Running in Production\n\nFor production deployments:\n\n```bash\nnpm install\nnpm run build\nnpm run preview\n```\n\nThe server listens on `http://localhost:4173` by default.\n" }, { "path": "entrypoint.sh", "content": "ENV_LOCAL_PATH=/app/.env.local\n\nif test -z \"${DOTENV_LOCAL}\" ; then\n if ! test -f \"${ENV_LOCAL_PATH}\" ; then\n echo \"DOTENV_LOCAL was not found in the ENV variables and .env.local is not set using a bind volume. Make sure to set environment variables properly. \"\n fi;\nelse\n echo \"DOTENV_LOCAL was found in the ENV variables. Creating .env.local file.\"\n cat <<< \"$DOTENV_LOCAL\" > ${ENV_LOCAL_PATH}\nfi;\n\nif [ \"$INCLUDE_DB\" = \"true\" ] ; then\n echo \"Starting local MongoDB instance\"\n nohup mongod &\nfi;\n\nexport PUBLIC_VERSION=$(node -p \"require('./package.json').version\")\n\ndotenv -e /app/.env -c -- node --dns-result-order=ipv4first /app/build/index.js -- --host 0.0.0.0 --port 3000" }, { "path": "models/add-your-models-here.txt", "content": "You can add .gguf files to this folder, and they will be picked up automatically by chat-ui. " }, { "path": "package.json", "content": "{\n\t\"name\": \"chat-ui\",\n\t\"version\": \"0.20.0\",\n\t\"private\": true,\n\t\"packageManager\": \"npm@9.5.0\",\n\t\"scripts\": {\n\t\t\"dev\": \"vite dev\",\n\t\t\"build\": \"vite build\",\n\t\t\"build:static\": \"ADAPTER=static vite build\",\n\t\t\"preview\": \"vite preview\",\n\t\t\"check\": \"svelte-kit sync && svelte-check --tsconfig ./tsconfig.json\",\n\t\t\"check:watch\": \"svelte-kit sync && svelte-check --tsconfig ./tsconfig.json --watch\",\n\t\t\"lint\": \"prettier --check . && eslint .\",\n\t\t\"format\": \"prettier --write .\",\n\t\t\"test\": \"vitest\",\n\t\t\"updateLocalEnv\": \"vite-node --options.transformMode.ssr='/.*/' scripts/updateLocalEnv.ts\",\n\t\t\"populate\": \"vite-node --options.transformMode.ssr='/.*/' scripts/populate.ts\",\n\t\t\"config\": \"vite-node --options.transformMode.ssr='/.*/' scripts/config.ts\",\n\t\t\"prepare\": \"husky\"\n\t},\n\t\"devDependencies\": {\n\t\t\"@faker-js/faker\": \"^8.4.1\",\n\t\t\"@iconify-json/carbon\": \"^1.1.16\",\n\t\t\"@iconify-json/eos-icons\": \"^1.1.6\",\n\t\t\"@iconify-json/lucide\": \"^1.2.77\",\n\t\t\"@sveltejs/adapter-node\": \"^5.2.12\",\n\t\t\"@sveltejs/adapter-static\": \"^3.0.8\",\n\t\t\"@sveltejs/kit\": \"^2.52.2\",\n\t\t\"@sveltejs/vite-plugin-svelte\": \"^5.0.3\",\n\t\t\"@tailwindcss/typography\": \"^0.5.9\",\n\t\t\"@types/dompurify\": \"^3.0.5\",\n\t\t\"@types/js-yaml\": \"^4.0.9\",\n\t\t\"@types/katex\": \"^0.16.7\",\n\t\t\"@types/mime-types\": \"^2.1.4\",\n\t\t\"@types/minimist\": \"^1.2.5\",\n\t\t\"@types/node\": \"^22.1.0\",\n\t\t\"@types/parquetjs\": \"^0.10.3\",\n\t\t\"@types/uuid\": \"^9.0.8\",\n\t\t\"@types/yazl\": \"^3.3.0\",\n\t\t\"@typescript-eslint/eslint-plugin\": \"^6.x\",\n\t\t\"@typescript-eslint/parser\": \"^6.x\",\n\t\t\"bson-objectid\": \"^2.0.4\",\n\t\t\"dompurify\": \"^3.2.4\",\n\t\t\"eslint\": \"^8.28.0\",\n\t\t\"eslint-config-prettier\": \"^8.5.0\",\n\t\t\"eslint-plugin-svelte\": \"^2.45.1\",\n\t\t\"husky\": \"^9.0.11\",\n\t\t\"isomorphic-dompurify\": \"2.13.0\",\n\t\t\"js-yaml\": \"^4.1.1\",\n\t\t\"lint-staged\": \"^15.2.7\",\n\t\t\"minimist\": \"^1.2.8\",\n\t\t\"mongodb-memory-server\": \"^10.1.2\",\n\t\t\"playwright\": \"^1.55.1\",\n\t\t\"prettier\": \"^3.5.3\",\n\t\t\"prettier-plugin-svelte\": \"^3.2.6\",\n\t\t\"prettier-plugin-tailwindcss\": \"^0.6.11\",\n\t\t\"sade\": \"^1.8.1\",\n\t\t\"superjson\": \"^2.2.2\",\n\t\t\"svelte\": \"^5.53.7\",\n\t\t\"svelte-check\": \"^4.0.0\",\n\t\t\"tslib\": \"^2.4.1\",\n\t\t\"typescript\": \"^5.5.0\",\n\t\t\"unplugin-icons\": \"^0.16.1\",\n\t\t\"vite\": \"^6.3.5\",\n\t\t\"vite-node\": \"^3.0.9\",\n\t\t\"vitest\": \"^3.1.4\",\n\t\t\"vitest-browser-svelte\": \"^0.1.0\",\n\t\t\"yazl\": \"^3.3.1\"\n\t},\n\t\"type\": \"module\",\n\t\"dependencies\": {\n\t\t\"@huggingface/hub\": \"^2.2.0\",\n\t\t\"@huggingface/inference\": \"^4.11.3\",\n\t\t\"@iconify-json/bi\": \"^1.1.21\",\n\t\t\"@modelcontextprotocol/sdk\": \"^1.26.0\",\n\t\t\"@resvg/resvg-js\": \"^2.6.2\",\n\t\t\"ajv\": \"^8.18.0\",\n\t\t\"autoprefixer\": \"^10.4.14\",\n\t\t\"bits-ui\": \"^2.14.2\",\n\t\t\"date-fns\": \"^2.29.3\",\n\t\t\"devalue\": \"^5.6.4\",\n\t\t\"dotenv\": \"^16.5.0\",\n\t\t\"file-type\": \"^21.3.1\",\n\t\t\"handlebars\": \"^4.7.8\",\n\t\t\"highlight.js\": \"^11.7.0\",\n\t\t\"htmlparser2\": \"^10.0.0\",\n\t\t\"ip-address\": \"^9.0.5\",\n\t\t\"jsdom\": \"^28.1.0\",\n\t\t\"json5\": \"^2.2.3\",\n\t\t\"katex\": \"^0.16.21\",\n\t\t\"marked\": \"^12.0.1\",\n\t\t\"mime-types\": \"^2.1.35\",\n\t\t\"mongodb\": \"^5.8.0\",\n\t\t\"nanoid\": \"^5.0.9\",\n\t\t\"openai\": \"^4.44.0\",\n\t\t\"openid-client\": \"^5.4.2\",\n\t\t\"parquetjs\": \"^0.11.2\",\n\t\t\"pino\": \"^9.0.0\",\n\t\t\"pino-pretty\": \"^11.0.0\",\n\t\t\"postcss\": \"^8.4.31\",\n\t\t\"prom-client\": \"^15.1.3\",\n\t\t\"qs\": \"^6.14.2\",\n\t\t\"satori\": \"^0.10.11\",\n\t\t\"satori-html\": \"^0.3.2\",\n\t\t\"sharp\": \"^0.33.4\",\n\t\t\"tailwind-scrollbar\": \"^3.0.0\",\n\t\t\"tailwindcss\": \"^3.4.0\",\n\t\t\"undici\": \"^7.18.2\",\n\t\t\"uuid\": \"^10.0.0\",\n\t\t\"web-haptics\": \"^0.0.6\",\n\t\t\"zod\": \"^3.22.3\"\n\t},\n\t\"overrides\": {\n\t\t\"@reflink/reflink\": \"file:stub/@reflink/reflink\"\n\t}\n}\n" }, { "path": "postcss.config.js", "content": "export default {\n\tplugins: {\n\t\ttailwindcss: {},\n\t\tautoprefixer: {},\n\t},\n};\n" }, { "path": "scripts/config.ts", "content": "import sade from \"sade\";\n\n// @ts-expect-error: vite-node makes the var available but the typescript compiler doesn't see them\nimport { config, ready } from \"$lib/server/config\";\n\nconst prog = sade(\"config\");\nawait ready;\nprog\n\t.command(\"clear\")\n\t.describe(\"Clear all config keys\")\n\t.action(async () => {\n\t\tconsole.log(\"Clearing config...\");\n\t\tawait clear();\n\t});\n\nprog\n\t.command(\"add \")\n\t.describe(\"Add a new config key\")\n\t.action(async (key: string, value: string) => {\n\t\tawait add(key, value);\n\t});\n\nprog\n\t.command(\"remove \")\n\t.describe(\"Remove a config key\")\n\t.action(async (key: string) => {\n\t\tconsole.log(`Removing ${key}`);\n\t\tawait remove(key);\n\t\tprocess.exit(0);\n\t});\n\nprog\n\t.command(\"help\")\n\t.describe(\"Show help information\")\n\t.action(() => {\n\t\tprog.help();\n\t\tprocess.exit(0);\n\t});\n\nasync function clear() {\n\tawait config.clear();\n\tprocess.exit(0);\n}\n\nasync function add(key: string, value: string) {\n\tif (!key || !value) {\n\t\tconsole.error(\"Key and value are required\");\n\t\tprocess.exit(1);\n\t}\n\tawait config.set(key as keyof typeof config.keysFromEnv, value);\n\tprocess.exit(0);\n}\n\nasync function remove(key: string) {\n\tif (!key) {\n\t\tconsole.error(\"Key is required\");\n\t\tprocess.exit(1);\n\t}\n\tawait config.delete(key as keyof typeof config.keysFromEnv);\n\tprocess.exit(0);\n}\n\n// Parse arguments and handle help automatically\nprog.parse(process.argv);\n" }, { "path": "scripts/populate.ts", "content": "import readline from \"readline\";\nimport minimist from \"minimist\";\n\n// @ts-expect-error: vite-node makes the var available but the typescript compiler doesn't see them\nimport { env } from \"$env/dynamic/private\";\n\nimport { faker } from \"@faker-js/faker\";\nimport { ObjectId } from \"mongodb\";\n\n// @ts-expect-error: vite-node makes the var available but the typescript compiler doesn't see them\nimport { ready } from \"$lib/server/config\";\nimport { collections } from \"$lib/server/database.ts\";\nimport { models } from \"../src/lib/server/models.ts\";\nimport type { User } from \"../src/lib/types/User\";\nimport type { Assistant } from \"../src/lib/types/Assistant\";\nimport type { Conversation } from \"../src/lib/types/Conversation\";\nimport type { Settings } from \"../src/lib/types/Settings\";\nimport { Message } from \"../src/lib/types/Message.ts\";\n\nimport { addChildren } from \"../src/lib/utils/tree/addChildren.ts\";\nimport { generateSearchTokens } from \"../src/lib/utils/searchTokens.ts\";\nimport { ReviewStatus } from \"../src/lib/types/Review.ts\";\nimport fs from \"fs\";\nimport path from \"path\";\n\nconst rl = readline.createInterface({\n\tinput: process.stdin,\n\toutput: process.stdout,\n});\n\nawait ready;\n\nrl.on(\"close\", function () {\n\tprocess.exit(0);\n});\n\nconst samples = fs.readFileSync(path.join(__dirname, \"samples.txt\"), \"utf8\").split(\"\\n---\\n\");\n\nconst possibleFlags = [\"reset\", \"all\", \"users\", \"settings\", \"assistants\", \"conversations\"];\nconst argv = minimist(process.argv.slice(2));\nconst flags = argv[\"_\"].filter((flag) => possibleFlags.includes(flag));\n\nasync function generateMessages(preprompt?: string): Promise {\n\tconst isLinear = faker.datatype.boolean(0.5);\n\tconst isInterrupted = faker.datatype.boolean(0.05);\n\n\tconst messages: Message[] = [];\n\n\tmessages.push({\n\t\tid: crypto.randomUUID(),\n\t\tfrom: \"system\",\n\t\tcontent: preprompt ?? \"\",\n\t\tcreatedAt: faker.date.recent({ days: 30 }),\n\t\tupdatedAt: faker.date.recent({ days: 30 }),\n\t});\n\n\tlet isUser = true;\n\tlet lastId = messages[0].id;\n\tif (isLinear) {\n\t\tconst convLength = faker.number.int({ min: 1, max: 25 }) * 2; // must always be even\n\n\t\tfor (let i = 0; i < convLength; i++) {\n\t\t\tlastId = addChildren(\n\t\t\t\t{\n\t\t\t\t\tmessages,\n\t\t\t\t\trootMessageId: messages[0].id,\n\t\t\t\t},\n\t\t\t\t{\n\t\t\t\t\tfrom: isUser ? \"user\" : \"assistant\",\n\t\t\t\t\tcontent:\n\t\t\t\t\t\tfaker.lorem.sentence({\n\t\t\t\t\t\t\tmin: 10,\n\t\t\t\t\t\t\tmax: isUser ? 50 : 200,\n\t\t\t\t\t\t}) +\n\t\t\t\t\t\t(!isUser && Math.random() < 0.1\n\t\t\t\t\t\t\t? \"\\n```\\n\" + faker.helpers.arrayElement(samples) + \"\\n```\\n\"\n\t\t\t\t\t\t\t: \"\"),\n\t\t\t\t\tcreatedAt: faker.date.recent({ days: 30 }),\n\t\t\t\t\tupdatedAt: faker.date.recent({ days: 30 }),\n\t\t\t\t\tinterrupted: !isUser && i === convLength - 1 && isInterrupted,\n\t\t\t\t},\n\t\t\t\tlastId\n\t\t\t);\n\t\t\tisUser = !isUser;\n\t\t}\n\t} else {\n\t\tconst convLength = faker.number.int({ min: 2, max: 200 });\n\n\t\tfor (let i = 0; i < convLength; i++) {\n\t\t\taddChildren(\n\t\t\t\t{\n\t\t\t\t\tmessages,\n\t\t\t\t\trootMessageId: messages[0].id,\n\t\t\t\t},\n\t\t\t\t{\n\t\t\t\t\tfrom: isUser ? \"user\" : \"assistant\",\n\t\t\t\t\tcontent:\n\t\t\t\t\t\tfaker.lorem.sentence({\n\t\t\t\t\t\t\tmin: 10,\n\t\t\t\t\t\t\tmax: isUser ? 50 : 200,\n\t\t\t\t\t\t}) +\n\t\t\t\t\t\t(!isUser && Math.random() < 0.1\n\t\t\t\t\t\t\t? \"\\n```\\n\" + faker.helpers.arrayElement(samples) + \"\\n```\\n\"\n\t\t\t\t\t\t\t: \"\"),\n\t\t\t\t\tcreatedAt: faker.date.recent({ days: 30 }),\n\t\t\t\t\tupdatedAt: faker.date.recent({ days: 30 }),\n\t\t\t\t\tinterrupted: !isUser && i === convLength - 1 && isInterrupted,\n\t\t\t\t},\n\t\t\t\tfaker.helpers.arrayElement([\n\t\t\t\t\tmessages[0].id,\n\t\t\t\t\t...messages.filter((m) => m.from === (isUser ? \"assistant\" : \"user\")).map((m) => m.id),\n\t\t\t\t])\n\t\t\t);\n\n\t\t\tisUser = !isUser;\n\t\t}\n\t}\n\treturn messages;\n}\n\nasync function seed() {\n\tconsole.log(\"Seeding...\");\n\tconst modelIds = models.map((model) => model.id);\n\n\tif (flags.includes(\"reset\")) {\n\t\tconsole.log(\"Starting reset of DB\");\n\t\tawait collections.users.deleteMany({});\n\t\tawait collections.settings.deleteMany({});\n\t\tawait collections.assistants.deleteMany({});\n\t\tawait collections.conversations.deleteMany({});\n\t\tawait collections.migrationResults.deleteMany({});\n\t\tawait collections.semaphores.deleteMany({});\n\t\tconsole.log(\"Reset done\");\n\t}\n\n\tif (flags.includes(\"users\") || flags.includes(\"all\")) {\n\t\tconsole.log(\"Creating 100 new users\");\n\t\tconst newUsers: User[] = Array.from({ length: 100 }, () => ({\n\t\t\t_id: new ObjectId(),\n\t\t\tcreatedAt: faker.date.recent({ days: 30 }),\n\t\t\tupdatedAt: faker.date.recent({ days: 30 }),\n\t\t\tusername: faker.internet.userName(),\n\t\t\tname: faker.person.fullName(),\n\t\t\thfUserId: faker.string.alphanumeric(24),\n\t\t\tavatarUrl: faker.image.avatar(),\n\t\t}));\n\n\t\tawait collections.users.insertMany(newUsers);\n\t\tconsole.log(\"Done creating users.\");\n\t}\n\n\tconst users = await collections.users.find().toArray();\n\tif (flags.includes(\"settings\") || flags.includes(\"all\")) {\n\t\tconsole.log(\"Updating settings for all users\");\n\t\tusers.forEach(async (user) => {\n\t\t\tconst settings: Settings = {\n\t\t\t\tuserId: user._id,\n\t\t\t\tshareConversationsWithModelAuthors: faker.datatype.boolean(0.25),\n\t\t\t\thideEmojiOnSidebar: faker.datatype.boolean(0.25),\n\t\t\t\tactiveModel: faker.helpers.arrayElement(modelIds),\n\t\t\t\tcreatedAt: faker.date.recent({ days: 30 }),\n\t\t\t\tupdatedAt: faker.date.recent({ days: 30 }),\n\t\t\t\tdisableStream: faker.datatype.boolean(0.25),\n\t\t\t\tdirectPaste: faker.datatype.boolean(0.25),\n\t\t\t\thidePromptExamples: {},\n\t\t\t\tcustomPrompts: {},\n\t\t\t\tassistants: [],\n\t\t\t};\n\t\t\tawait collections.settings.updateOne(\n\t\t\t\t{ userId: user._id },\n\t\t\t\t{ $set: { ...settings } },\n\t\t\t\t{ upsert: true }\n\t\t\t);\n\t\t});\n\t\tconsole.log(\"Done updating settings.\");\n\t}\n\n\tif (flags.includes(\"assistants\") || flags.includes(\"all\")) {\n\t\tconsole.log(\"Creating assistants for all users\");\n\t\tawait Promise.all(\n\t\t\tusers.map(async (user) => {\n\t\t\t\tconst name = faker.animal.insect();\n\t\t\t\tconst assistants = faker.helpers.multiple(\n\t\t\t\t\t() => ({\n\t\t\t\t\t\t_id: new ObjectId(),\n\t\t\t\t\t\tname,\n\t\t\t\t\t\tcreatedById: user._id,\n\t\t\t\t\t\tcreatedByName: user.username,\n\t\t\t\t\t\tcreatedAt: faker.date.recent({ days: 30 }),\n\t\t\t\t\t\tupdatedAt: faker.date.recent({ days: 30 }),\n\t\t\t\t\t\tuserCount: faker.number.int({ min: 1, max: 100000 }),\n\t\t\t\t\t\treview: faker.helpers.enumValue(ReviewStatus),\n\t\t\t\t\t\tmodelId: faker.helpers.arrayElement(modelIds),\n\t\t\t\t\t\tdescription: faker.lorem.sentence(),\n\t\t\t\t\t\tpreprompt: faker.hacker.phrase(),\n\t\t\t\t\t\texampleInputs: faker.helpers.multiple(() => faker.lorem.sentence(), {\n\t\t\t\t\t\t\tcount: faker.number.int({ min: 0, max: 4 }),\n\t\t\t\t\t\t}),\n\t\t\t\t\t\tsearchTokens: generateSearchTokens(name),\n\t\t\t\t\t\tlast24HoursCount: faker.number.int({ min: 0, max: 1000 }),\n\t\t\t\t\t}),\n\t\t\t\t\t{ count: faker.number.int({ min: 3, max: 10 }) }\n\t\t\t\t);\n\t\t\t\tawait collections.assistants.insertMany(assistants);\n\t\t\t\tawait collections.settings.updateOne(\n\t\t\t\t\t{ userId: user._id },\n\t\t\t\t\t{ $set: { assistants: assistants.map((a) => a._id.toString()) } },\n\t\t\t\t\t{ upsert: true }\n\t\t\t\t);\n\t\t\t})\n\t\t);\n\t\tconsole.log(\"Done creating assistants.\");\n\t}\n\n\tif (flags.includes(\"conversations\") || flags.includes(\"all\")) {\n\t\tconsole.log(\"Creating conversations for all users\");\n\t\tawait Promise.all(\n\t\t\tusers.map(async (user) => {\n\t\t\t\tconst conversations = faker.helpers.multiple(\n\t\t\t\t\tasync () => {\n\t\t\t\t\t\tconst settings = await collections.settings.findOne({ userId: user._id });\n\n\t\t\t\t\t\tconst assistantId =\n\t\t\t\t\t\t\tsettings?.assistants && settings.assistants.length > 0 && faker.datatype.boolean(0.1)\n\t\t\t\t\t\t\t\t? faker.helpers.arrayElement(settings.assistants)\n\t\t\t\t\t\t\t\t: undefined;\n\n\t\t\t\t\t\tconst preprompt =\n\t\t\t\t\t\t\t(assistantId\n\t\t\t\t\t\t\t\t? await collections.assistants\n\t\t\t\t\t\t\t\t\t\t.findOne({ _id: assistantId })\n\t\t\t\t\t\t\t\t\t\t.then((assistant: Assistant) => assistant?.preprompt ?? \"\")\n\t\t\t\t\t\t\t\t: faker.helpers.maybe(() => faker.hacker.phrase(), { probability: 0.5 })) ?? \"\";\n\n\t\t\t\t\t\tconst messages = await generateMessages(preprompt);\n\n\t\t\t\t\t\tconst conv = {\n\t\t\t\t\t\t\t_id: new ObjectId(),\n\t\t\t\t\t\t\tuserId: user._id,\n\t\t\t\t\t\t\tassistantId,\n\t\t\t\t\t\t\tpreprompt,\n\t\t\t\t\t\t\tcreatedAt: faker.date.recent({ days: 145 }),\n\t\t\t\t\t\t\tupdatedAt: faker.date.recent({ days: 145 }),\n\t\t\t\t\t\t\tmodel: faker.helpers.arrayElement(modelIds),\n\t\t\t\t\t\t\ttitle: faker.internet.emoji() + \" \" + faker.hacker.phrase(),\n\t\t\t\t\t\t\t// embeddings removed in this build\n\t\t\t\t\t\t\tmessages,\n\t\t\t\t\t\t\trootMessageId: messages[0].id,\n\t\t\t\t\t\t} satisfies Conversation;\n\n\t\t\t\t\t\treturn conv;\n\t\t\t\t\t},\n\t\t\t\t\t{ count: faker.number.int({ min: 10, max: 200 }) }\n\t\t\t\t);\n\n\t\t\t\tawait collections.conversations.insertMany(await Promise.all(conversations));\n\t\t\t})\n\t\t);\n\t\tconsole.log(\"Done creating conversations.\");\n\t}\n}\n\n// run seed\n(async () => {\n\ttry {\n\t\trl.question(\n\t\t\t\"You're about to run a seeding script on the following MONGODB_URL: \\x1b[31m\" +\n\t\t\t\tenv.MONGODB_URL +\n\t\t\t\t\"\\x1b[0m\\n\\n With the following flags: \\x1b[31m\" +\n\t\t\t\tflags.join(\"\\x1b[0m , \\x1b[31m\") +\n\t\t\t\t\"\\x1b[0m\\n \\n\\n Are you sure you want to continue? (yes/no): \",\n\t\t\tasync (confirm) => {\n\t\t\t\tif (confirm !== \"yes\") {\n\t\t\t\t\tconsole.log(\"Not 'yes', exiting.\");\n\t\t\t\t\trl.close();\n\t\t\t\t\tprocess.exit(0);\n\t\t\t\t}\n\t\t\t\tconsole.log(\"Starting seeding...\");\n\t\t\t\tawait seed();\n\t\t\t\tconsole.log(\"Seeding done.\");\n\t\t\t\trl.close();\n\t\t\t}\n\t\t);\n\t} catch (e) {\n\t\tconsole.error(e);\n\t\tprocess.exit(1);\n\t}\n})();\n" }, { "path": "scripts/samples.txt", "content": "import { Observable, of, from, interval, throwError } from 'rxjs';\nimport { map, filter, catchError, switchMap, take, tap } from 'rxjs/operators';\n\n// Mock function to fetch stock prices (simulates API call)\nconst fetchStockPrice = (ticker: string): Observable => {\n return new Observable((observer) => {\n const intervalId = setInterval(() => {\n if (Math.random() < 0.1) { // Simulating an error 10% of the time\n observer.error(`Error fetching stock price for ${ticker}`);\n } else {\n const price = parseFloat((Math.random() * 1000).toFixed(2));\n observer.next(price);\n }\n }, 1000);\n\n return () => {\n clearInterval(intervalId);\n console.log(`Stopped fetching prices for ${ticker}`);\n };\n });\n};\n\n// Example usage: Tracking stock price updates\nconst stockTicker = 'AAPL';\nconst stockPrice$ = fetchStockPrice(stockTicker).pipe(\n map(price => ({ ticker: stockTicker, price })), // Transform data\n filter(data => data.price > 500), // Only keep prices above 500\n tap(data => console.log(`Price update:`, data)), // Side effect: Logging\n catchError(err => {\n console.error(err);\n return of({ ticker: stockTicker, price: null }); // Fallback observable\n })\n);\n\n// Subscribe to the stock price updates\nconst subscription = stockPrice$.subscribe({\n next: data => console.log(`Subscriber received:`, data),\n error: err => console.error(`Subscription error:`, err),\n complete: () => console.log('Stream complete'),\n});\n\n// Automatically unsubscribe after 10 seconds\nsetTimeout(() => {\n subscription.unsubscribe();\n console.log('Unsubscribed from stock price updates.');\n}, 10000);\n---\nclass EnforceAttrsMeta(type):\n \"\"\"\n Metaclass that enforces the presence of specific attributes in a class\n and automatically decorates methods with a logging wrapper.\n \"\"\"\n \n required_attributes = ['name', 'version']\n\n def __new__(cls, name, bases, class_dict):\n \"\"\"\n Create a new class with enforced attributes and method logging.\n\n :param name: Name of the class being created.\n :param bases: Tuple of base classes.\n :param class_dict: Dictionary of attributes and methods of the class.\n :return: Newly created class object.\n \"\"\"\n # Ensure required attributes exist\n for attr in cls.required_attributes:\n if attr not in class_dict:\n raise TypeError(f\"Class '{name}' is missing required attribute '{attr}'\")\n\n # Wrap all methods in a logging decorator\n for key, value in class_dict.items():\n if callable(value): # Check if it's a method\n class_dict[key] = cls.log_calls(value)\n\n return super().__new__(cls, name, bases, class_dict)\n\n @staticmethod\n def log_calls(func):\n \"\"\"\n Decorator that logs method calls and arguments.\n\n :param func: Function to be wrapped.\n :return: Wrapped function with logging.\n \"\"\"\n def wrapper(*args, **kwargs):\n print(f\"Calling {func.__name__} with args={args} kwargs={kwargs}\")\n result = func(*args, **kwargs)\n print(f\"{func.__name__} returned {result}\")\n return result\n return wrapper\n\n\nclass PluginBase(metaclass=EnforceAttrsMeta):\n \"\"\"\n Base class for plugins that enforces required attributes and logging.\n \"\"\"\n name = \"BasePlugin\"\n version = \"1.0\"\n\n def run(self, data):\n \"\"\"\n Process the input data.\n\n :param data: The data to be processed.\n :return: Processed result.\n \"\"\"\n return f\"Processed {data}\"\n\n\nclass CustomPlugin(PluginBase):\n \"\"\"\n Custom plugin that extends PluginBase and adheres to enforced rules.\n \"\"\"\n name = \"CustomPlugin\"\n version = \"2.0\"\n\n def run(self, data):\n \"\"\"\n Custom processing logic.\n\n :param data: The data to process.\n :return: Modified data.\n \"\"\"\n return f\"Custom processing of {data}\"\n\n\n# Uncommenting the following class definition will raise a TypeError\n# because 'version' attribute is missing.\n# class InvalidPlugin(PluginBase):\n# name = \"InvalidPlugin\"\n\n\nif __name__ == \"__main__\":\n # Instantiate and use the plugin\n plugin = CustomPlugin()\n print(plugin.run(\"example data\"))\n---\n\n\n\n \n \n Click the Box Game\n \n\n\n

Click the Box!

\n

Score: 0

\n
\n
\n
\n \n\n\n" }, { "path": "scripts/setups/vitest-setup-client.ts", "content": "" }, { "path": "scripts/setups/vitest-setup-server.ts", "content": "import { vi, afterAll } from \"vitest\";\nimport dotenv from \"dotenv\";\nimport { resolve } from \"path\";\nimport fs from \"fs\";\nimport { MongoMemoryServer } from \"mongodb-memory-server\";\n\nlet mongoServer: MongoMemoryServer;\n// Load the .env file\nconst envPath = resolve(__dirname, \"../../.env\");\ndotenv.config({ path: envPath });\n\n// Read the .env file content\nconst envContent = fs.readFileSync(envPath, \"utf-8\");\n\n// Parse the .env content\nconst envVars = dotenv.parse(envContent);\n\n// Separate public and private variables\nconst publicEnv = {};\nconst privateEnv = {};\n\nfor (const [key, value] of Object.entries(envVars)) {\n\tif (key.startsWith(\"PUBLIC_\")) {\n\t\tpublicEnv[key] = value;\n\t} else {\n\t\tprivateEnv[key] = value;\n\t}\n}\n\nvi.mock(\"$env/dynamic/public\", () => ({\n\tenv: publicEnv,\n}));\n\nvi.mock(\"$env/dynamic/private\", async () => {\n\tmongoServer = await MongoMemoryServer.create();\n\n\treturn {\n\t\tenv: {\n\t\t\t...privateEnv,\n\t\t\tMONGODB_URL: mongoServer.getUri(),\n\t\t},\n\t};\n});\n\nafterAll(async () => {\n\tif (mongoServer) {\n\t\tawait mongoServer.stop();\n\t}\n});\n" }, { "path": "scripts/updateLocalEnv.ts", "content": "import fs from \"fs\";\nimport yaml from \"js-yaml\";\n\nconst file = fs.readFileSync(\"chart/env/prod.yaml\", \"utf8\");\n\n// have to do a weird stringify/parse because of some node error\nconst prod = JSON.parse(JSON.stringify(yaml.load(file)));\nconst vars = prod.envVars as Record;\n\nlet PUBLIC_CONFIG = \"\";\n\nObject.entries(vars)\n\t// filter keys used in prod with the proxy\n\t.filter(\n\t\t([key]) =>\n\t\t\t![\n\t\t\t\t\"XFF_DEPTH\",\n\t\t\t\t\"ADDRESS_HEADER\",\n\t\t\t\t\"APP_BASE\",\n\t\t\t\t\"PUBLIC_ORIGIN\",\n\t\t\t\t\"PUBLIC_SHARE_PREFIX\",\n\t\t\t\t\"ADMIN_CLI_LOGIN\",\n\t\t\t].includes(key)\n\t)\n\t.forEach(([key, value]) => {\n\t\tPUBLIC_CONFIG += `${key}=\\`${value}\\`\\n`;\n\t});\n\nconst SECRET_CONFIG =\n\t(fs.existsSync(\".env.SECRET_CONFIG\")\n\t\t? fs.readFileSync(\".env.SECRET_CONFIG\", \"utf8\")\n\t\t: process.env.SECRET_CONFIG) ?? \"\";\n\n// Prepend the content of the env variable SECRET_CONFIG\nlet full_config = `${PUBLIC_CONFIG}\\n${SECRET_CONFIG}`;\n\n// replace the internal proxy url with the public endpoint\nfull_config = full_config.replaceAll(\n\t\"https://internal.api-inference.huggingface.co\",\n\t\"https://router.huggingface.co/hf-inference\"\n);\n\nfull_config = full_config.replaceAll(\"COOKIE_SECURE=`true`\", \"COOKIE_SECURE=`false`\");\nfull_config = full_config.replaceAll(\"LOG_LEVEL=`debug`\", \"LOG_LEVEL=`info`\");\nfull_config = full_config.replaceAll(\"NODE_ENV=`prod`\", \"NODE_ENV=`development`\");\n\n// Write full_config to .env.local\nfs.writeFileSync(\".env.local\", full_config);\n" }, { "path": "server.log", "content": "/Users/vm/.venv/bin/python3: No module named uvicorn\n/Users/vm/.venv/bin/python3: No module named uvicorn\n" }, { "path": "src/ambient.d.ts", "content": "declare module \"*.ttf\" {\n\tconst value: ArrayBuffer;\n\texport default value;\n}\n\n// Legacy helpers removed: web search support is deprecated, so we intentionally\n// avoid leaking those shapes into the global ambient types.\n" }, { "path": "src/app.d.ts", "content": "/// \n/// \n\nimport type { User } from \"$lib/types/User\";\n\n// See https://kit.svelte.dev/docs/types#app\n// for information about these interfaces\ndeclare global {\n\tnamespace App {\n\t\t// interface Error {}\n\t\tinterface Locals {\n\t\t\tsessionId: string;\n\t\t\tuser?: User;\n\t\t\tisAdmin: boolean;\n\t\t\ttoken?: string;\n\t\t\t/** Organization to bill inference requests to (from settings) */\n\t\t\tbillingOrganization?: string;\n\t\t}\n\n\t\tinterface Error {\n\t\t\tmessage: string;\n\t\t\terrorId?: ReturnType;\n\t\t}\n\t\t// interface PageData {}\n\t\t// interface Platform {}\n\t}\n}\n\nexport {};\n" }, { "path": "src/app.html", "content": "\n\n\t\n\t\t\n\t\t\n\t\t\n\t\t\n\t\t%sveltekit.head%\n\t\n\t\n\t\t
%sveltekit.body%
\n\n\t\t\n\t\t\n\t\n\n" }, { "path": "src/hooks.server.ts", "content": "import { building } from \"$app/environment\";\nimport type { Handle, HandleServerError, ServerInit, HandleFetch } from \"@sveltejs/kit\";\nimport { initServer } from \"$lib/server/hooks/init\";\nimport { handleRequest } from \"$lib/server/hooks/handle\";\nimport { handleServerError } from \"$lib/server/hooks/error\";\nimport { handleFetchRequest } from \"$lib/server/hooks/fetch\";\n\nexport const init: ServerInit = async () => {\n\tif (building) return;\n\treturn initServer();\n};\n\nexport const handle: Handle = async (input) => {\n\tif (building) {\n\t\t// During static build, still replace %gaId% placeholder with empty string\n\t\t// to prevent the GA script from loading with an invalid ID\n\t\treturn input.resolve(input.event, {\n\t\t\ttransformPageChunk: ({ html }) => html.replace(\"%gaId%\", \"\"),\n\t\t});\n\t}\n\treturn handleRequest(input);\n};\n\nexport const handleError: HandleServerError = async (input) => {\n\tif (building) throw input.error;\n\treturn handleServerError(input);\n};\n\nexport const handleFetch: HandleFetch = async (input) => {\n\tif (building) return input.fetch(input.request);\n\treturn handleFetchRequest(input);\n};\n" }, { "path": "src/hooks.ts", "content": "import { publicConfigTransporter } from \"$lib/utils/PublicConfig.svelte\";\nimport type { Transport } from \"@sveltejs/kit\";\n\nexport const transport: Transport = {\n\tPublicConfig: publicConfigTransporter,\n};\n" }, { "path": "src/lib/APIClient.ts", "content": "import { base } from \"$app/paths\";\nimport { browser } from \"$app/environment\";\nimport superjson from \"superjson\";\nimport ObjectId from \"bson-objectid\";\n\nsuperjson.registerCustom(\n\t{\n\t\tisApplicable: (value): value is ObjectId => {\n\t\t\tif (typeof value !== \"string\" && ObjectId.isValid(value)) {\n\t\t\t\tconst str = value.toString();\n\t\t\t\treturn /^[0-9a-fA-F]{24}$/.test(str);\n\t\t\t}\n\t\t\treturn false;\n\t\t},\n\t\tserialize: (value) => value.toString(),\n\t\tdeserialize: (value) => new ObjectId(value),\n\t},\n\t\"ObjectId\"\n);\n\ntype FetchFn = typeof globalThis.fetch;\n\ninterface ApiResponse {\n\tdata: T | null;\n\terror: unknown;\n\tstatus: number;\n}\n\nasync function apiCall(\n\tfetcher: FetchFn,\n\turl: string,\n\tmethod: string,\n\tbody?: unknown,\n\tquery?: Record\n): Promise> {\n\tconst u = new URL(url);\n\tif (query) {\n\t\tfor (const [k, v] of Object.entries(query)) {\n\t\t\tif (v !== undefined && v !== null) {\n\t\t\t\tu.searchParams.set(k, String(v));\n\t\t\t}\n\t\t}\n\t}\n\n\tconst init: RequestInit = { method };\n\tif (body !== undefined && body !== null) {\n\t\tinit.headers = { \"Content-Type\": \"application/json\" };\n\t\tinit.body = JSON.stringify(body);\n\t}\n\n\tconst res = await fetcher(u.toString(), init);\n\tif (!res.ok) {\n\t\tlet errorBody: unknown;\n\t\ttry {\n\t\t\terrorBody = await res.json();\n\t\t} catch {\n\t\t\terrorBody = await res.text().catch(() => res.statusText);\n\t\t}\n\t\treturn { data: null, error: errorBody, status: res.status };\n\t}\n\n\t// Handle empty responses (e.g. POST /user/settings returns empty body)\n\tconst text = await res.text();\n\tif (!text) {\n\t\treturn { data: null, error: null, status: res.status };\n\t}\n\n\treturn { data: text as unknown as T, error: null, status: res.status };\n}\n\nfunction endpoint(fetcher: FetchFn, baseUrl: string) {\n\treturn {\n\t\tget(opts?: { query?: Record }) {\n\t\t\treturn apiCall(fetcher, baseUrl, \"GET\", undefined, opts?.query);\n\t\t},\n\t\tpost(body?: unknown) {\n\t\t\treturn apiCall(fetcher, baseUrl, \"POST\", body);\n\t\t},\n\t\tpatch(body?: unknown) {\n\t\t\treturn apiCall(fetcher, baseUrl, \"PATCH\", body);\n\t\t},\n\t\tdelete() {\n\t\t\treturn apiCall(fetcher, baseUrl, \"DELETE\");\n\t\t},\n\t};\n}\n\nexport function useAPIClient({\n\tfetch: customFetch,\n\torigin,\n}: {\n\tfetch?: FetchFn;\n\torigin?: string;\n} = {}) {\n\tconst fetcher = customFetch ?? globalThis.fetch;\n\tconst baseUrl = browser\n\t\t? `${window.location.origin}${base}/api/v2`\n\t\t: `${origin ?? `http://localhost:5173`}${base}/api/v2`;\n\n\treturn {\n\t\tconversations: Object.assign(\n\t\t\t// client.conversations({ id: \"...\" }) — returns endpoint for /conversations/:id\n\t\t\t(params: { id: string }) => ({\n\t\t\t\t...endpoint(fetcher, `${baseUrl}/conversations/${params.id}`),\n\t\t\t\tmessage: (msgParams: { messageId: string }) =>\n\t\t\t\t\tendpoint(fetcher, `${baseUrl}/conversations/${params.id}/message/${msgParams.messageId}`),\n\t\t\t}),\n\t\t\t// client.conversations.get(), .delete()\n\t\t\t{\n\t\t\t\t...endpoint(fetcher, `${baseUrl}/conversations`),\n\t\t\t\t\"import-share\": endpoint(fetcher, `${baseUrl}/conversations/import-share`),\n\t\t\t}\n\t\t),\n\t\tuser: {\n\t\t\t...endpoint(fetcher, `${baseUrl}/user`),\n\t\t\tsettings: endpoint(fetcher, `${baseUrl}/user/settings`),\n\t\t\treports: endpoint(fetcher, `${baseUrl}/user/reports`),\n\t\t\t\"billing-orgs\": endpoint(fetcher, `${baseUrl}/user/billing-orgs`),\n\t\t},\n\t\tmodels: {\n\t\t\t...endpoint(fetcher, `${baseUrl}/models`),\n\t\t\told: endpoint(fetcher, `${baseUrl}/models/old`),\n\t\t\trefresh: endpoint(fetcher, `${baseUrl}/models/refresh`),\n\t\t},\n\t\t\"public-config\": endpoint(fetcher, `${baseUrl}/public-config`),\n\t\t\"feature-flags\": endpoint(fetcher, `${baseUrl}/feature-flags`),\n\t\tdebug: {\n\t\t\tconfig: endpoint(fetcher, `${baseUrl}/debug/config`),\n\t\t\trefresh: endpoint(fetcher, `${baseUrl}/debug/refresh`),\n\t\t},\n\t\texport: endpoint(fetcher, `${baseUrl}/export`),\n\t};\n}\n\n// eslint-disable-next-line @typescript-eslint/no-explicit-any\nexport function handleResponse(response: ApiResponse): any {\n\tif (response.error) {\n\t\tthrow new Error(JSON.stringify(response.error));\n\t}\n\n\tif (response.data === null) {\n\t\treturn null;\n\t}\n\n\treturn superjson.parse(\n\t\ttypeof response.data === \"string\" ? response.data : JSON.stringify(response.data)\n\t);\n}\n" }, { "path": "src/lib/actions/clickOutside.ts", "content": "export function clickOutside(element: HTMLElement, callbackFunction: () => void) {\n\tfunction onClick(event: MouseEvent) {\n\t\tif (!element.contains(event.target as Node)) {\n\t\t\tcallbackFunction();\n\t\t}\n\t}\n\n\tdocument.body.addEventListener(\"click\", onClick);\n\n\treturn {\n\t\tupdate(newCallbackFunction: () => void) {\n\t\t\tcallbackFunction = newCallbackFunction;\n\t\t},\n\t\tdestroy() {\n\t\t\tdocument.body.removeEventListener(\"click\", onClick);\n\t\t},\n\t};\n}\n" }, { "path": "src/lib/actions/snapScrollToBottom.ts", "content": "import { navigating } from \"$app/state\";\nimport { tick } from \"svelte\";\n\n// Threshold to determine if user is \"at bottom\" - larger value prevents false detachment\nconst BOTTOM_THRESHOLD = 50;\nconst USER_SCROLL_DEBOUNCE_MS = 150;\nconst PROGRAMMATIC_SCROLL_GRACE_MS = 100;\nconst TOUCH_DETACH_THRESHOLD_PX = 10;\n\ninterface ScrollDependency {\n\tsignal: unknown;\n\tforceReattach?: number;\n}\n\ntype MaybeScrollDependency = ScrollDependency | unknown;\n\nconst getForceReattach = (value: MaybeScrollDependency): number => {\n\tif (typeof value === \"object\" && value !== null && \"forceReattach\" in value) {\n\t\treturn (value as ScrollDependency).forceReattach ?? 0;\n\t}\n\treturn 0;\n};\n\n/**\n * Auto-scroll action that snaps to bottom while respecting user scroll intent.\n *\n * Key behaviors:\n * 1. Uses wheel/touch events to detect actual user intent\n * 2. Uses IntersectionObserver on a sentinel element to reliably detect \"at bottom\" state\n * 3. Larger threshold to prevent edge-case false detachments\n *\n * @param node element to snap scroll to bottom\n * @param dependency pass in { signal, forceReattach } - signal triggers scroll updates,\n * forceReattach (counter) forces re-attachment when incremented\n */\nexport const snapScrollToBottom = (node: HTMLElement, dependency: MaybeScrollDependency) => {\n\t// --- State ----------------------------------------------------------------\n\n\t// Track whether user has intentionally scrolled away from bottom\n\tlet isDetached = false;\n\n\t// Track the last forceReattach value to detect changes\n\tlet lastForceReattach = getForceReattach(dependency);\n\n\t// Track if user is actively scrolling (via wheel/touch)\n\tlet userScrolling = false;\n\tlet userScrollTimeout: ReturnType | undefined;\n\n\t// Track programmatic scrolls to avoid treating them as user scrolls\n\tlet isProgrammaticScroll = false;\n\tlet lastProgrammaticScrollTime = 0;\n\n\t// Track previous scroll position to detect scrollbar drags\n\tlet prevScrollTop = node.scrollTop;\n\n\t// Touch handling state\n\tlet touchStartY = 0;\n\n\t// Observers and sentinel\n\tlet resizeObserver: ResizeObserver | undefined;\n\tlet intersectionObserver: IntersectionObserver | undefined;\n\tlet sentinel: HTMLDivElement | undefined;\n\n\t// Track content height for early-return optimization during streaming\n\tlet lastScrollHeight = node.scrollHeight;\n\n\t// --- Helpers --------------------------------------------------------------\n\n\tconst clearUserScrollTimeout = () => {\n\t\tif (userScrollTimeout) {\n\t\t\tclearTimeout(userScrollTimeout);\n\t\t\tuserScrollTimeout = undefined;\n\t\t}\n\t};\n\n\tconst distanceFromBottom = () => node.scrollHeight - node.scrollTop - node.clientHeight;\n\n\tconst isAtBottom = () => distanceFromBottom() <= BOTTOM_THRESHOLD;\n\n\tconst scrollToBottom = () => {\n\t\tisProgrammaticScroll = true;\n\t\tlastProgrammaticScrollTime = Date.now();\n\n\t\tnode.scrollTo({ top: node.scrollHeight });\n\n\t\tif (typeof requestAnimationFrame === \"function\") {\n\t\t\trequestAnimationFrame(() => {\n\t\t\t\tisProgrammaticScroll = false;\n\t\t\t});\n\t\t} else {\n\t\t\tisProgrammaticScroll = false;\n\t\t}\n\t};\n\n\tconst settleScrollAfterLayout = async () => {\n\t\tif (typeof requestAnimationFrame !== \"function\") return;\n\n\t\tconst raf = () => new Promise((resolve) => requestAnimationFrame(() => resolve()));\n\n\t\tawait raf();\n\t\tif (!userScrolling && !isDetached) {\n\t\t\tscrollToBottom();\n\t\t}\n\n\t\tawait raf();\n\t\tif (!userScrolling && !isDetached) {\n\t\t\tscrollToBottom();\n\t\t}\n\t};\n\n\tconst scheduleUserScrollEndCheck = () => {\n\t\tuserScrolling = true;\n\t\tclearUserScrollTimeout();\n\n\t\tuserScrollTimeout = setTimeout(() => {\n\t\t\tuserScrolling = false;\n\n\t\t\t// If user scrolled back to bottom, re-attach\n\t\t\tif (isAtBottom()) {\n\t\t\t\tisDetached = false;\n\t\t\t}\n\n\t\t\t// Re-trigger scroll if still attached, to catch content that arrived during scrolling\n\t\t\tif (!isDetached) {\n\t\t\t\tscrollToBottom();\n\t\t\t}\n\t\t}, USER_SCROLL_DEBOUNCE_MS);\n\t};\n\n\tconst createSentinel = () => {\n\t\tsentinel = document.createElement(\"div\");\n\t\tsentinel.style.height = \"1px\";\n\t\tsentinel.style.width = \"100%\";\n\t\tsentinel.setAttribute(\"aria-hidden\", \"true\");\n\t\tsentinel.setAttribute(\"data-scroll-sentinel\", \"\");\n\n\t\t// Find the content container (first child) and append sentinel there\n\t\tconst container = node.firstElementChild;\n\t\tif (container) {\n\t\t\tcontainer.appendChild(sentinel);\n\t\t} else {\n\t\t\tnode.appendChild(sentinel);\n\t\t}\n\t};\n\n\tconst setupIntersectionObserver = () => {\n\t\tif (typeof IntersectionObserver === \"undefined\" || !sentinel) return;\n\n\t\tintersectionObserver = new IntersectionObserver(\n\t\t\t(entries) => {\n\t\t\t\tconst entry = entries[0];\n\n\t\t\t\t// If sentinel is visible and user isn't actively scrolling, we're at bottom\n\t\t\t\tif (entry?.isIntersecting && !userScrolling) {\n\t\t\t\t\tisDetached = false;\n\t\t\t\t\t// Immediately scroll to catch up with any content that arrived while detached\n\t\t\t\t\tscrollToBottom();\n\t\t\t\t}\n\t\t\t},\n\t\t\t{\n\t\t\t\troot: node,\n\t\t\t\tthreshold: 0,\n\t\t\t\trootMargin: `0px 0px ${BOTTOM_THRESHOLD}px 0px`,\n\t\t\t}\n\t\t);\n\n\t\tintersectionObserver.observe(sentinel);\n\t};\n\n\tconst setupResizeObserver = () => {\n\t\tif (typeof ResizeObserver === \"undefined\") return;\n\n\t\tconst target = node.firstElementChild ?? node;\n\t\tresizeObserver = new ResizeObserver(() => {\n\t\t\t// Don't auto-scroll if user has detached and we're not navigating\n\t\t\tif (isDetached && !navigating.to) return;\n\t\t\t// Don't interrupt active user scrolling\n\t\t\tif (userScrolling) return;\n\n\t\t\tscrollToBottom();\n\t\t});\n\n\t\tresizeObserver.observe(target);\n\t};\n\n\t// --- Action update logic --------------------------------------------------\n\n\tconst handleForceReattach = async (newDependency: MaybeScrollDependency) => {\n\t\tconst forceReattach = getForceReattach(newDependency);\n\n\t\tif (forceReattach > lastForceReattach) {\n\t\t\tlastForceReattach = forceReattach;\n\t\t\tisDetached = false;\n\t\t\tuserScrolling = false;\n\t\t\tclearUserScrollTimeout();\n\n\t\t\tawait tick();\n\t\t\tscrollToBottom();\n\t\t\treturn true;\n\t\t}\n\n\t\treturn false;\n\t};\n\n\tasync function updateScroll(newDependency?: MaybeScrollDependency) {\n\t\t// 1. Explicit force re-attach\n\t\tif (newDependency && (await handleForceReattach(newDependency))) {\n\t\t\treturn;\n\t\t}\n\n\t\t// 2. Don't scroll if user has detached and we're not navigating\n\t\tif (isDetached && !navigating.to) return;\n\n\t\t// 3. Don't scroll if user is actively scrolling\n\t\tif (userScrolling) return;\n\n\t\t// 4. Early return if already at bottom and no content change (perf optimization for streaming)\n\t\tconst currentHeight = node.scrollHeight;\n\t\tif (isAtBottom() && currentHeight === lastScrollHeight) {\n\t\t\treturn;\n\t\t}\n\t\tlastScrollHeight = currentHeight;\n\n\t\t// 5. Wait for DOM to update, then scroll and settle after layout shifts\n\t\tawait tick();\n\t\tscrollToBottom();\n\t\tawait settleScrollAfterLayout();\n\t}\n\n\t// --- Event handlers -------------------------------------------------------\n\n\t// Detect user scroll intent via wheel events (mouse/trackpad)\n\tconst handleWheel = (event: WheelEvent) => {\n\t\tconst { deltaY } = event;\n\n\t\t// User is scrolling up - detach\n\t\tif (deltaY < 0) {\n\t\t\tisDetached = true;\n\t\t}\n\n\t\t// User is scrolling down - check for re-attachment immediately\n\t\t// This ensures fast re-attachment when user scrolls to bottom during fast generation\n\t\tif (deltaY > 0 && isAtBottom()) {\n\t\t\tisDetached = false;\n\t\t\tuserScrolling = false;\n\t\t\tclearUserScrollTimeout();\n\t\t\tscrollToBottom();\n\t\t\treturn;\n\t\t}\n\n\t\tscheduleUserScrollEndCheck();\n\t};\n\n\t// Detect user scroll intent via touch events (mobile)\n\tconst handleTouchStart = (event: TouchEvent) => {\n\t\ttouchStartY = event.touches[0]?.clientY ?? 0;\n\t};\n\n\tconst handleTouchMove = (event: TouchEvent) => {\n\t\tconst touchY = event.touches[0]?.clientY ?? 0;\n\t\tconst deltaY = touchStartY - touchY;\n\n\t\t// User is scrolling up (finger moving down)\n\t\tif (deltaY < -TOUCH_DETACH_THRESHOLD_PX) {\n\t\t\tisDetached = true;\n\t\t}\n\n\t\t// User is scrolling down (finger moving up) - check for re-attachment immediately\n\t\tif (deltaY > TOUCH_DETACH_THRESHOLD_PX && isAtBottom()) {\n\t\t\tisDetached = false;\n\t\t\tuserScrolling = false;\n\t\t\tclearUserScrollTimeout();\n\t\t\tscrollToBottom();\n\t\t\ttouchStartY = touchY;\n\t\t\treturn;\n\t\t}\n\n\t\tscheduleUserScrollEndCheck();\n\t\ttouchStartY = touchY;\n\t};\n\n\t// Handle scroll events to detect scrollbar usage and re-attach when at bottom\n\tconst handleScroll = () => {\n\t\tconst now = Date.now();\n\t\tconst timeSinceLastProgrammaticScroll = now - lastProgrammaticScrollTime;\n\t\tconst inGracePeriod =\n\t\t\tisProgrammaticScroll || timeSinceLastProgrammaticScroll < PROGRAMMATIC_SCROLL_GRACE_MS;\n\n\t\t// If not from wheel/touch, this is likely a scrollbar drag\n\t\tif (!userScrolling) {\n\t\t\tconst scrollingUp = node.scrollTop < prevScrollTop;\n\n\t\t\t// Always allow detach (scrolling up) - don't ignore user intent\n\t\t\tif (scrollingUp) {\n\t\t\t\tisDetached = true;\n\t\t\t}\n\n\t\t\t// Only re-attach when at bottom if NOT in grace period\n\t\t\t// (avoids false re-attach from content resize pushing scroll position)\n\t\t\tif (!inGracePeriod && isAtBottom()) {\n\t\t\t\tisDetached = false;\n\t\t\t\t// Immediately scroll to catch up with any content that arrived while detached\n\t\t\t\tscrollToBottom();\n\t\t\t}\n\t\t}\n\n\t\tprevScrollTop = node.scrollTop;\n\t};\n\n\t// --- Setup ----------------------------------------------------------------\n\n\tnode.addEventListener(\"wheel\", handleWheel, { passive: true });\n\tnode.addEventListener(\"touchstart\", handleTouchStart, { passive: true });\n\tnode.addEventListener(\"touchmove\", handleTouchMove, { passive: true });\n\tnode.addEventListener(\"scroll\", handleScroll, { passive: true });\n\n\tcreateSentinel();\n\tsetupIntersectionObserver();\n\tsetupResizeObserver();\n\n\t// Initial scroll if we have content\n\tif (dependency) {\n\t\tvoid (async () => {\n\t\t\tawait tick();\n\t\t\tscrollToBottom();\n\t\t})();\n\t}\n\n\t// --- Cleanup --------------------------------------------------------------\n\n\treturn {\n\t\tupdate: updateScroll,\n\t\tdestroy: () => {\n\t\t\tclearUserScrollTimeout();\n\n\t\t\tnode.removeEventListener(\"wheel\", handleWheel);\n\t\t\tnode.removeEventListener(\"touchstart\", handleTouchStart);\n\t\t\tnode.removeEventListener(\"touchmove\", handleTouchMove);\n\t\t\tnode.removeEventListener(\"scroll\", handleScroll);\n\n\t\t\tresizeObserver?.disconnect();\n\t\t\tintersectionObserver?.disconnect();\n\t\t\tsentinel?.remove();\n\t\t},\n\t};\n};\n" }, { "path": "src/lib/buildPrompt.ts", "content": "import type { EndpointParameters } from \"./server/endpoints/endpoints\";\nimport type { BackendModel } from \"./server/models\";\n\ntype buildPromptOptions = Pick & {\n\tmodel: BackendModel;\n};\n\nexport async function buildPrompt({\n\tmessages,\n\tmodel,\n\tpreprompt,\n}: buildPromptOptions): Promise {\n\tconst filteredMessages = messages;\n\n\tif (filteredMessages[0].from === \"system\" && preprompt) {\n\t\tfilteredMessages[0].content = preprompt;\n\t}\n\n\tconst prompt = model\n\t\t.chatPromptRender({\n\t\t\tmessages: filteredMessages.map((m) => ({\n\t\t\t\t...m,\n\t\t\t\trole: m.from,\n\t\t\t})),\n\t\t\tpreprompt,\n\t\t})\n\t\t// Not super precise, but it's truncated in the model's backend anyway\n\t\t.split(\" \")\n\t\t.slice(-(model.parameters?.truncate ?? 0))\n\t\t.join(\" \");\n\n\treturn prompt;\n}\n" }, { "path": "src/lib/components/AnnouncementBanner.svelte", "content": "\n\n
\n\tNew\n\t{title}\n\t
\n\t\t{@render children?.()}\n\t
\n
\n" }, { "path": "src/lib/components/BackgroundGenerationPoller.svelte", "content": "\n" }, { "path": "src/lib/components/CodeBlock.svelte", "content": "\n\n
\n\t
\n\t\t\n\t\t\t{#if showPreview}\n\t\t\t\t {\n\t\t\t\t\t\tif (!loading) {\n\t\t\t\t\t\t\tpreviewOpen = true;\n\t\t\t\t\t\t}\n\t\t\t\t\t}}\n\t\t\t\t\ttitle=\"Preview HTML\"\n\t\t\t\t\taria-label=\"Preview HTML\"\n\t\t\t\t>\n\t\t\t\t\t{#if loading}\n\t\t\t\t\t\t\n\t\t\t\t\t{:else}\n\t\t\t\t\t\t\n\t\t\t\t\t{/if}\n\t\t\t\t\tPreview\n\t\t\t\t\n\t\t\t{/if}\n\t\t\t\n\t\t
\n\t
\n\t
{@html DOMPurify.sanitize(code)}
\n\n\t{#if previewOpen}\n\t\t (previewOpen = false)} />\n\t{/if}\n\n" }, { "path": "src/lib/components/CopyToClipBoardBtn.svelte", "content": "\n\n {\n\t\tonClick?.();\n\t\thandleClick();\n\t}}\n>\n\t
\n\t\t{#if children}{@render children()}{:else}\n\t\t\t\n\t\t{/if}\n\n\t\t{#if showTooltip}\n\t\t\t\n\t\t{/if}\n\t
\n\n" }, { "path": "src/lib/components/DeleteConversationModal.svelte", "content": "\n\n{#if open}\n\t\n\t\t
\n\t\t\t
\n\t\t\t\t

Delete conversation

\n\t\t\t\t\n\t\t\t
\n\n\t\t\t

\n\t\t\t\tAre you sure you want to delete \"{title}\"? This action\n\t\t\t\tcannot be undone.\n\t\t\t

\n\n\t\t\t
\n\t\t\t\t\n\t\t\t\t\tCancel\n\t\t\t\t\n\t\t\t\t\n\t\t\t\t\tDelete\n\t\t\t\t\n\t\t\t
\n\t\t
\n\t\n{/if}\n" }, { "path": "src/lib/components/EditConversationModal.svelte", "content": "\n\n{#if open}\n\t\n\t\t {\n\t\t\t\te.preventDefault();\n\t\t\t\tsave();\n\t\t\t}}\n\t\t>\n\t\t\t
\n\t\t\t\t

Rename conversation

\n\t\t\t\t\n\t\t\t
\n\n\t\t\t
\n\t\t\t\t\n\t\t\t\t (newTitle = (e.currentTarget as HTMLInputElement).value)}\n\t\t\t\t\tclass=\"w-full rounded-xl border border-gray-200 bg-white px-3 py-2 text-[15px] text-gray-800 outline-none placeholder:text-gray-400 focus:ring-2 focus:ring-gray-200 dark:border-gray-700 dark:bg-gray-800 dark:text-gray-100 dark:placeholder:text-gray-500 dark:focus:ring-gray-700\"\n\t\t\t\t\tplaceholder=\"Enter a title\"\n\t\t\t\t/>\n\t\t\t
\n\n\t\t\t
\n\t\t\t\t\n\t\t\t\t\tCancel\n\t\t\t\t\n\t\t\t\t\n\t\t\t\t\tSave\n\t\t\t\t\n\t\t\t
\n\t\t\n\t\n{/if}\n" }, { "path": "src/lib/components/ExpandNavigation.svelte", "content": "\n\n\n\t
\n\t
\n\n" }, { "path": "src/lib/components/HoverTooltip.svelte", "content": "\n\n
\n\t{@render children?.()}\n\n\t\n\t\t{label}\n\t
\n\n" }, { "path": "src/lib/components/HtmlPreviewModal.svelte", "content": "\n\n\n\n onclose?.()}\n>\n\t
\n\t\t\n\n\t\t\n\t\t onclose?.()}\n\t\t>\n\t\t\t\n\t\t\tClose preview\n\t\t\n\n\t\t{#if errors.length > 0}\n\t\t\t {\n\t\t\t\t\tpendingChatInput.set(composeText());\n\t\t\t\t\tonclose?.();\n\t\t\t\t}}\n\t\t\t>\n\t\t\t\tError caught ({errors.length})\n\t\t\t\n\t\t{/if}\n\t
\n\n" }, { "path": "src/lib/components/InfiniteScroll.svelte", "content": "\n\n
\n" }, { "path": "src/lib/components/MobileNav.svelte", "content": "\n\n\n\n\n\t (isOpen = true)}\n\t\taria-label=\"Open menu\"\n\t\tbind:this={openEl}>\n\t
\n\t\t{#if page.params?.id}\n\t\t\t{title}\n\t\t{/if}\n\t
\n\t
\n\t\t{#if isHuggingChat}\n\t\t\t {\n\t\t\t\t\tif (!canShare) return;\n\t\t\t\t\tshareModal.open();\n\t\t\t\t}}\n\t\t\t\taria-label=\"Share conversation\"\n\t\t\t>\n\t\t\t\t\n\t\t\t\n\t\t{/if}\n\t\t {\n\t\t\t\tif (requireAuthUser()) {\n\t\t\t\t\te.preventDefault();\n\t\t\t\t}\n\t\t\t}}\n\t\t>\n\t\t\t\n\t\t\n\t
\n\n\n\n{#if isOpen || isDragging}\n\t\n{/if}\n\n\n\t{@render children?.()}\n\n" }, { "path": "src/lib/components/Modal.svelte", "content": "\n\n\n\t {\n\t\t\te.stopPropagation();\n\t\t\thandleBackdropClick(e);\n\t\t}}\n\t\ttransition:fade|local={{ easing: cubicOut, duration: 300 }}\n\t\tclass=\"fixed inset-0 z-40 flex items-center justify-center bg-black/80 backdrop-blur-sm dark:bg-black/50\"\n\t>\n\t\t{#if disableFly}\n\t\t\t\n\t\t\t\t{#if closeButton}\n\t\t\t\t\t\n\t\t\t\t{/if}\n\t\t\t\t{@render children?.()}\n\t\t\t\n\t\t{:else}\n\t\t\t\n\t\t\t\t{#if closeButton}\n\t\t\t\t\t\n\t\t\t\t{/if}\n\t\t\t\t{@render children?.()}\n\t\t\t\n\t\t{/if}\n\t\n\n" }, { "path": "src/lib/components/ModelCardMetadata.svelte", "content": "\n\n\n\t\n\t\tModel\n\t\t
 page
\n\t{#if model.datasetName || model.datasetUrl}\n\t\t\n\t\t\tDataset\n\t\t\t
 page
\n\t{/if}\n\t{#if model.hasInferenceAPI}\n\t\t\n\t\t\tAPI\n\t\t\n\t{/if}\n\t{#if model.websiteUrl}\n\t\t\n\t\t\t{#if model.name.startsWith(\"meta-llama/Meta-Llama\")}\n\t\t\t\t\n\t\t\t\tBuilt with Llama\n\t\t\t{:else}\n\t\t\t\t\n\t\t\t\tWebsite\n\t\t\t{/if}\n\t\t\n\t{/if}\n\n" }, { "path": "src/lib/components/NavConversationItem.svelte", "content": "\n\n {\n\t\tif (e.detail >= 2) {\n\t\t\te.preventDefault();\n\t\t\tstartInlineEdit();\n\t\t}\n\t}}\n>\n\t{#if inlineEditing}\n\t\t (inlineTitle = (e.currentTarget as HTMLInputElement).value)}\n\t\t\tonkeydown={(e) => {\n\t\t\t\tif (e.key === \"Enter\") {\n\t\t\t\t\te.preventDefault();\n\t\t\t\t\tcommitInlineEdit();\n\t\t\t\t} else if (e.key === \"Escape\") {\n\t\t\t\t\te.preventDefault();\n\t\t\t\t\tcancelInlineEdit();\n\t\t\t\t}\n\t\t\t}}\n\t\t\tonblur={commitInlineEdit}\n\t\t\tonclick={(e) => e.preventDefault()}\n\t\t\tclass=\"my-0 h-full min-w-0 flex-1 truncate border-none bg-transparent p-0 text-inherit outline-none first-letter:uppercase focus:ring-0\"\n\t\t/>\n\t{:else}\n\t\t
\n\t\t\t{conv.title}\n\t\t
\n\t{/if}\n\n\t{#if !readOnly && !inlineEditing}\n\t\t {\n\t\t\t\te.preventDefault();\n\t\t\t\tif (requireAuthUser()) return;\n\t\t\t\trenameOpen = true;\n\t\t\t}}\n\t\t>\n\t\t\t\n\t\t\n\n\t\t {\n\t\t\t\tevent.preventDefault();\n\t\t\t\tif (requireAuthUser()) return;\n\t\t\t\tif (event.shiftKey) {\n\t\t\t\t\tondeleteConversation?.(conv.id.toString());\n\t\t\t\t} else {\n\t\t\t\t\tdeleteOpen = true;\n\t\t\t\t}\n\t\t\t}}\n\t\t>\n\t\t\t\n\t\t\n\t{/if}\n\n\n\n{#if renameOpen}\n\t (renameOpen = false)}\n\t\tonsave={(payload) => {\n\t\t\trenameOpen = false;\n\t\t\toneditConversationTitle?.({ id: conv.id.toString(), title: payload.title });\n\t\t}}\n\t/>\n{/if}\n\n\n{#if deleteOpen}\n\t (deleteOpen = false)}\n\t\tondelete={() => {\n\t\t\tdeleteOpen = false;\n\t\t\tondeleteConversation?.(conv.id.toString());\n\t\t}}\n\t/>\n{/if}\n" }, { "path": "src/lib/components/NavMenu.svelte", "content": "\n\n\n\n\n\t\n\t\t\n\t\t{publicConfig.PUBLIC_APP_NAME}\n\t\n\t\n\t\tNew Chat\n\t\n\n\n\n\t
\n\t\t{#each Object.entries(groupedConversations) as [group, convs]}\n\t\t\t{#if convs.length}\n\t\t\t\t

\n\t\t\t\t\t{titles[group]}\n\t\t\t\t

\n\t\t\t\t{#each convs as conv}\n\t\t\t\t\t\n\t\t\t\t{/each}\n\t\t\t{/if}\n\t\t{/each}\n\t
\n\t{#if hasMore}\n\t\t\n\t{/if}\n\n\n\t{#if user?.username || user?.email}\n\t\t\n\t\t\t\n\t\t\t{user?.username || user?.email}\n\n\t\t\t{#if publicConfig.isHuggingChat && $isPro === false}\n\t\t\t\t\n\t\t\t\t\t\n\t\t\t\t\tGet PRO\n\t\t\t\t\n\t\t\t{:else if publicConfig.isHuggingChat && $isPro === true}\n\t\t\t\t\n\t\t\t\t\t\n\t\t\t\t\tPRO\n\t\t\t\t\n\t\t\t{/if}\n\t\t\n\t{/if}\n\t\n\t\tModels\n\t\t{nModels}\n\t\n\n\t{#if user?.username || user?.email}\n\t\t (showMcpModal = true)}\n\t\t\tclass=\"flex h-9 flex-none items-center gap-1.5 rounded-lg pl-2.5 pr-2 text-gray-500 hover:bg-gray-100 dark:text-gray-400 dark:hover:bg-gray-700\"\n\t\t>\n\t\t\tMCP Servers\n\t\t\t{#if $enabledServersCount > 0}\n\t\t\t\t\n\t\t\t\t\t{$enabledServersCount}\n\t\t\t\t\n\t\t\t{/if}\n\t\t\n\t{/if}\n\n\t\n\t\t\n\t\t\tSettings\n\t\t\n\t\t {\n\t\t\t\tswitchTheme();\n\t\t\t}}\n\t\t\taria-label=\"Toggle theme\"\n\t\t\tclass=\"flex size-9 min-w-[1.5em] flex-none items-center justify-center rounded-lg p-2 text-gray-500 hover:bg-gray-100 dark:text-gray-400 dark:hover:bg-gray-700\"\n\t\t>\n\t\t\t{#if browser}\n\t\t\t\t{#if isDark}\n\t\t\t\t\t\n\t\t\t\t{:else}\n\t\t\t\t\t\n\t\t\t\t{/if}\n\t\t\t{/if}\n\t\t\n\t\n\n\n{#if showMcpModal}\n\t (showMcpModal = false)} />\n{/if}\n" }, { "path": "src/lib/components/Pagination.svelte", "content": "\n\n{#if numTotalPages > 1}\n\t\n{/if}\n" }, { "path": "src/lib/components/PaginationArrow.svelte", "content": "\n\n\n\t{#if direction === \"previous\"}\n\t\t\n\t\tPrevious\n\t{:else}\n\t\tNext\n\t\t\n\t{/if}\n\n" }, { "path": "src/lib/components/Portal.svelte", "content": "\n\n\n" }, { "path": "src/lib/components/RetryBtn.svelte", "content": "\n\n\n\t Retry\n\n" }, { "path": "src/lib/components/ScrollToBottomBtn.svelte", "content": "\n\n{#if visible}\n\t scrollNode.scrollTo({ top: scrollNode.scrollHeight, behavior: \"smooth\" })}\n\t\tclass=\"btn absolute flex h-[41px] w-[41px] rounded-full border bg-white shadow-md transition-all hover:bg-gray-100 dark:border-gray-600 dark:bg-gray-700 dark:shadow-gray-950 dark:hover:bg-gray-600 {className}\"\n\t\t>\n{/if}\n" }, { "path": "src/lib/components/ScrollToPreviousBtn.svelte", "content": "\n\n{#if visible}\n\t\n\t\t\n\t\n{/if}\n" }, { "path": "src/lib/components/ShareConversationModal.svelte", "content": "\n\n{#if open}\n\t\n\t\t
\n\t\t\t\n\t\t\t{#if createdUrl}\n\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t\t\t\tPublic link created\n\t\t\t\t\t
\n\t\t\t\t\t\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\tA public link to your chat has been created.\n\t\t\t\t
\n\t\t\t{:else}\n\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t\t\t\tShare public link to chat\n\t\t\t\t\t
\n\t\t\t\t\t\n\t\t\t\t
\n\t\t\t\t
\n\t\t\t\t\tAny messages you add after sharing stay private.\n\t\t\t\t
\n\t\t\t{/if}\n\n\t\t\t{#if errorMsg}\n\t\t\t\t\n\t\t\t\t\t{errorMsg}\n\t\t\t\t
\n\t\t\t{/if}\n\n\t\t\t\n\t\t\t\n\t\t\t\t\n\n\t\t\t\t{#if createdUrl}\n\t\t\t\t\t {\n\t\t\t\t\t\t\tjustCopied = true;\n\t\t\t\t\t\t\toncopied?.();\n\t\t\t\t\t\t\tsetTimeout(() => (justCopied = false), 1200);\n\t\t\t\t\t\t}}\n\t\t\t\t\t>\n\t\t\t\t\t\t{#snippet children()}\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t{#if justCopied}\n\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\tCopied\n\t\t\t\t\t\t\t\t{:else}\n\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\tCopy link\n\t\t\t\t\t\t\t\t{/if}\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t{/snippet}\n\t\t\t\t\t\n\t\t\t\t{:else}\n\t\t\t\t\t\n\t\t\t\t\t\t{#if creating}\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t\tCreating…\n\t\t\t\t\t\t{:else}\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t\tCreate link\n\t\t\t\t\t\t{/if}\n\t\t\t\t\t\n\t\t\t\t{/if}\n\t\t\t\n\t\t\n\t\n{/if}\n" }, { "path": "src/lib/components/StopGeneratingBtn.svelte", "content": "\n\n\n\tStop generating\n\t\n\n\n\n" }, { "path": "src/lib/components/SubscribeModal.svelte", "content": "\n\n\n\t\n\t\t\n\t\t\t
\n\t\t\t\t\n\t\t\t\t\t{#if $isPro}\n\t\t\t\t\t\t\n\t\t\t\t\t{:else}\n\t\t\t\t\t\t\n\t\t\t\t\t{/if}\n\t\t\t\t
\n\t\t\t\t

\n\t\t\t\t\t{$isPro ? \"Out of Credits\" : \"Upgrade Required\"}\n\t\t\t\t

\n\t\t\t\n\t\t\n\n\t\t
\n\t\t\t{#if $isPro}\n\t\t\t\t

\n\t\t\t\t\tYou've used all your available credits. Purchase additional credits to continue using\n\t\t\t\t\tHuggingChat.\n\t\t\t\t

\n\t\t\t\t

\n\t\t\t\t\tYour credits can be used in other HF services and external apps via Inference Providers.\n\t\t\t\t

\n\t\t\t{:else}\n\t\t\t\t

\n\t\t\t\t\tYou've reached your message limit. Upgrade to Hugging Face PRO to continue using\n\t\t\t\t\tHuggingChat.\n\t\t\t\t

\n\t\t\t\t

\n\t\t\t\t\tIt's also possible to use your PRO credits in your favorite AI tools.\n\t\t\t\t

\n\t\t\t{/if}\n\t\t
\n\n\t\t
\n\t\t\t{#if $isPro}\n\t\t\t\t\n\t\t\t\t\tPurchase Credits\n\t\t\t\t\n\t\t\t{:else}\n\t\t\t\t\n\t\t\t\t\tUpgrade to Pro\n\t\t\t\t\n\t\t\t{/if}\n\t\t\t\n\t\t\t\tMaybe later\n\t\t\t\n\t\t
\n\t\n\n" }, { "path": "src/lib/components/Switch.svelte", "content": "\n\n\ndiv]:translate-x-3.5\"\n>\n\t
\n\n" }, { "path": "src/lib/components/SystemPromptModal.svelte", "content": "\n\n (isOpen = !isOpen)}\n\tonkeydown={(e) => e.key === \"Enter\" && (isOpen = !isOpen)}\n>\n\t Using Custom System Prompt\n\n\n{#if isOpen}\n\t (isOpen = false)} width=\"w-full !max-w-xl\">\n\t\t
\n\t\t\t\n\t\t\t\t

System Prompt

\n\t\t\t\t\n\t\t\t
\n\t\t\t\n\t\t\n\t\n{/if}\n" }, { "path": "src/lib/components/Toast.svelte", "content": "\n\n\n\t\n\t\t\n\t\t\t\n\t\t\t

\n\t\t\t\t{message}\n\t\t\t

\n\t\t\n\t\n\n" }, { "path": "src/lib/components/Tooltip.svelte", "content": "\n\n\n\t\n\t{label}\n\n" }, { "path": "src/lib/components/WelcomeModal.svelte", "content": "\n\n\n\t\n\t\t\n\t\t\t\n\t\t\t\n\t\t\t\n\t\t\t\tNow with MCP!\n\t\t\t\n\t\t\n\n\t\t
\n\t\t\t

\n\t\t\t\tWelcome to {publicConfig.PUBLIC_APP_NAME}, the chat app powered by open source AI models.\n\t\t\t

\n\t\t\t

\n\t\t\t\t Omni automatically picks the best AI model to give\n\t\t\t\tyou optimal answers depending on your requests.\n\t\t\t

\n\t\t\t

\n\t\t\t\tYou can also choose from any available open source models to chat with directly.\n\t\t\t

\n\t\t
\n\n\t\t\n\t\t\tStart chatting\n\t\t\n\t\n\n" }, { "path": "src/lib/components/chat/Alternatives.svelte", "content": "\n\n\n\t onshowAlternateMsg?.({ id: alternatives[Math.max(0, currentIdx - 1)] })}\n\t\tdisabled={currentIdx === 0 || loading}\n\t>\n\t\t\n\t\n\t\n\t\t{currentIdx + 1} / {alternatives.length}\n\t\n\t\n\t\t\tonshowAlternateMsg?.({\n\t\t\t\tid: alternatives[Math.min(alternatives.length - 1, currentIdx + 1)],\n\t\t\t})}\n\t\tdisabled={currentIdx === alternatives.length - 1 || loading}\n\t>\n\t\t\n\t\n\t\n\n" }, { "path": "src/lib/components/chat/BlockWrapper.svelte", "content": "\n\n
\n\t\n\t
\n\t\t\n\t\t\t{@render icon()}\n\t\t\t{#if loading}\n\t\t\t\t\n\t\t\t\t\t\n\t\t\t\t\n\t\t\t{/if}\n\t\t
\n\t\t{#if hasNext}\n\t\t\t
\n\t\t{/if}\n\t
\n\n\t\n\t
\n\t\t{@render children()}\n\t
\n\n\n\n" }, { "path": "src/lib/components/chat/ChatInput.svelte", "content": "\n\n
\n\t (isCompositionOn = true)}\n\t\toncompositionend={() => (isCompositionOn = false)}\n\t\t{placeholder}\n\t\t{disabled}\n\t\tonfocus={handleFocus}\n\t\tonblur={handleBlur}\n\t\tonbeforeinput={requireAuthUser}\n\t>\n\n\t{#if !showNoTools}\n\t\t\n\t\t\t{#if showFileUpload}\n\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t\t\t{/if}\n\t\t\t\t
\n\t\t\t{/if}\n\t\t\n\t{/if}\n\t{@render children?.()}\n\n\t\n\n\t{#if isMcpManagerOpen}\n\t\t (isMcpManagerOpen = false)} />\n\t{/if}\n\n\n\n" }, { "path": "src/lib/components/chat/ChatIntroduction.svelte", "content": "\n\n
\n\t\n\t\t\n\t\t{publicConfig.PUBLIC_APP_NAME}\n\t
\n\t\n\n" }, { "path": "src/lib/components/chat/ChatMessage.svelte", "content": "\n\n{#if message.from === \"assistant\"}\n\t= messageWidth\n\t\t\t? 'mb-1'\n\t\t\t: ''}\"\n\t\tdata-message-id={message.id}\n\t\tdata-message-role=\"assistant\"\n\t\trole=\"presentation\"\n\t\tonclick={() => (isTapped = !isTapped)}\n\t\tonkeydown={() => (isTapped = !isTapped)}\n\t>\n\t\t\n\t\t\n\t\t\t{#if message.files?.length}\n\t\t\t\t
\n\t\t\t\t\t{#each message.files as file (file.value)}\n\t\t\t\t\t\t\n\t\t\t\t\t{/each}\n\t\t\t\t
\n\t\t\t{/if}\n\n\t\t\t\n\t\t\t
\n\t\t\t\t{#if isLast && loading && blocks.length === 0}\n\t\t\t\t\t\n\t\t\t\t{/if}\n\t\t\t\t{#each blocks as block, blockIndex (block.type === \"tool\" ? `${block.uuid}-${blockIndex}` : `text-${blockIndex}`)}\n\t\t\t\t\t{@const nextBlock = blocks[blockIndex + 1]}\n\t\t\t\t\t{@const nextBlockHasThink =\n\t\t\t\t\t\tnextBlock?.type === \"text\" && THINK_BLOCK_TEST_REGEX.test(nextBlock.content)}\n\t\t\t\t\t{@const nextIsLinkable = nextBlock?.type === \"tool\" || nextBlockHasThink}\n\t\t\t\t\t{#if block.type === \"tool\"}\n\t\t\t\t\t\t
\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t
\n\t\t\t\t\t{:else if block.type === \"text\"}\n\t\t\t\t\t\t{#if isLast && loading && block.content.length === 0}\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t{/if}\n\n\t\t\t\t\t\t{#if hasClientThink}\n\t\t\t\t\t\t\t{@const parts = block.content.split(THINK_BLOCK_REGEX)}\n\t\t\t\t\t\t\t{#each parts as part, partIndex}\n\t\t\t\t\t\t\t\t{@const remainingParts = parts.slice(partIndex + 1)}\n\t\t\t\t\t\t\t\t{@const hasMoreLinkable =\n\t\t\t\t\t\t\t\t\tremainingParts.some((p) => p && THINK_BLOCK_TEST_REGEX.test(p)) || nextIsLinkable}\n\t\t\t\t\t\t\t\t{#if part && part.startsWith(\"\")}\n\t\t\t\t\t\t\t\t\t{@const isClosed = part.endsWith(\"\")}\n\t\t\t\t\t\t\t\t\t{@const thinkContent = part.slice(7, isClosed ? -8 : undefined)}\n\n\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t{:else if part && part.trim().length > 0}\n\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\t
\n\t\t\t\t\t\t\t\t{/if}\n\t\t\t\t\t\t\t{/each}\n\t\t\t\t\t\t{:else}\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t{/if}\n\t\t\t\t\t{/if}\n\t\t\t\t{/each}\n\t\t\t\n\t\t\n\n\t\t{#if message.routerMetadata || (!loading && message.content)}\n\t\t\t messageWidth\n\t\t\t\t\t? 'left-1 pl-1 lg:pl-7'\n\t\t\t\t\t: 'right-1'} flex max-w-[calc(100dvw-40px)] items-center gap-0.5\"\n\t\t\t\tbind:offsetWidth={messageInfoWidth}\n\t\t\t>\n\t\t\t\t{#if message.routerMetadata && (message.routerMetadata.route || message.routerMetadata.model || message.routerMetadata.provider) && (!isLast || !loading)}\n\t\t\t\t\t\n\t\t\t\t\t\t{#if message.routerMetadata.route && message.routerMetadata.model}\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t{message.routerMetadata.route}\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t\twith\n\t\t\t\t\t\t\t{#if publicConfig.isHuggingChat}\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\t{message.routerMetadata.model.split(\"/\").pop()}\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t{:else}\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\t{message.routerMetadata.model.split(\"/\").pop()}\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t{/if}\n\t\t\t\t\t\t{/if}\n\t\t\t\t\t\t{#if message.routerMetadata.provider}\n\t\t\t\t\t\t\t{@const hubOrg = PROVIDERS_HUB_ORGS[message.routerMetadata.provider]}\n\t\t\t\t\t\t\tvia\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t ((e.currentTarget as HTMLImageElement).style.display = \"none\")}\n\t\t\t\t\t\t\t\t/>\n\t\t\t\t\t\t\t\t{message.routerMetadata.provider}\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t{/if}\n\t\t\t\t\t\n\t\t\t\t{/if}\n\t\t\t\t{#if !isLast || !loading}\n\t\t\t\t\t {\n\t\t\t\t\t\t\tisCopied = true;\n\t\t\t\t\t\t}}\n\t\t\t\t\t\tclassNames=\"btn rounded-sm p-1 text-sm text-gray-400 hover:text-gray-500 focus:ring-0 dark:text-gray-400 dark:hover:text-gray-300\"\n\t\t\t\t\t\tvalue={contentWithoutThink}\n\t\t\t\t\t\ticonClassNames=\"text-xs\"\n\t\t\t\t\t/>\n\t\t\t\t\t {\n\t\t\t\t\t\t\tonretry?.({ id: message.id });\n\t\t\t\t\t\t}}\n\t\t\t\t\t>\n\t\t\t\t\t\t\n\t\t\t\t\t\n\t\t\t\t\t{#if alternatives.length > 1 && editMsdgId === null}\n\t\t\t\t\t\t onshowAlternateMsg?.(payload)}\n\t\t\t\t\t\t/>\n\t\t\t\t\t{/if}\n\t\t\t\t{/if}\n\t\t\t\n\t\t{/if}\n\t\n\t{#if lightboxSrc}\n\t\t (lightboxSrc = null)} />\n\t{/if}\n{/if}\n{#if message.from === \"user\"}\n\t 1 && editMsdgId === null\n\t\t\t? 'mb-7'\n\t\t\t: ''} w-full items-start justify-start gap-4\"\n\t\tdata-message-id={message.id}\n\t\tdata-message-type=\"user\"\n\t\trole=\"presentation\"\n\t\tonclick={() => (isTapped = !isTapped)}\n\t\tonkeydown={() => (isTapped = !isTapped)}\n\t>\n\t\t
\n\t\t\t{#if message.files?.length}\n\t\t\t\t
\n\t\t\t\t\t{#each message.files as file}\n\t\t\t\t\t\t\n\t\t\t\t\t{/each}\n\t\t\t\t
\n\t\t\t{/if}\n\n\t\t\t
\n\t\t\t\t{#if !editMode}\n\t\t\t\t\t\n\t\t\t\t\t\t{message.content.trim()}\n\t\t\t\t\t

\n\t\t\t\t{:else}\n\t\t\t\t\t {\n\t\t\t\t\t\t\te.preventDefault();\n\t\t\t\t\t\t\tonretry?.({ content: editContentEl?.value, id: message.id });\n\t\t\t\t\t\t\teditMsdgId = null;\n\t\t\t\t\t\t}}\n\t\t\t\t\t>\n\t\t\t\t\t\t\n\t\t\t\t\t\t
\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\tSend\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t {\n\t\t\t\t\t\t\t\t\teditMsdgId = null;\n\t\t\t\t\t\t\t\t}}\n\t\t\t\t\t\t\t>\n\t\t\t\t\t\t\t\tCancel\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t
\n\t\t\t\t\t\n\t\t\t\t{/if}\n\t\t\t
\n\t\t\t
\n\t\t\t\t{#if alternatives.length > 1 && editMsdgId === null}\n\t\t\t\t\t onshowAlternateMsg?.(payload)}\n\t\t\t\t\t/>\n\t\t\t\t{/if}\n\t\t\t\t{#if (alternatives.length > 1 && editMsdgId === null) || (!loading && !editMode)}\n\t\t\t\t\t {\n\t\t\t\t\t\t\tif (requireAuthUser()) return;\n\t\t\t\t\t\t\teditMsdgId = message.id;\n\t\t\t\t\t\t}}\n\t\t\t\t\t>\n\t\t\t\t\t\t\n\t\t\t\t\t\tEdit\n\t\t\t\t\t\n\t\t\t\t{/if}\n\t\t\t
\n\t\t
\n\t\n{/if}\n\n\n" }, { "path": "src/lib/components/chat/ChatWindow.svelte", "content": "\n\n {\n\t\te.preventDefault();\n\t}}\n\tondrop={(e) => {\n\t\te.preventDefault();\n\t\tonDrag = false;\n\t}}\n/>\n\n
\n\t{#if shareModalOpen}\n\t\t shareModal.close()} />\n\t{/if}\n\t\n\t\t\n\t\t\t{#if preprompt && preprompt != currentModel.preprompt}\n\t\t\t\t\n\t\t\t{/if}\n\n\t\t\t{#if messages.length > 0}\n\t\t\t\t
\n\t\t\t\t\t{#each messages as message, idx (message.id)}\n\t\t\t\t\t\t a.includes(message.id)) ?? []}\n\t\t\t\t\t\t\tisAuthor={!shared}\n\t\t\t\t\t\t\treadOnly={isReadOnly}\n\t\t\t\t\t\t\tisLast={idx === messages.length - 1}\n\t\t\t\t\t\t\tbind:editMsdgId\n\t\t\t\t\t\t\tonretry={(payload) => onretry?.(payload)}\n\t\t\t\t\t\t\tonshowAlternateMsg={(payload) => onshowAlternateMsg?.(payload)}\n\t\t\t\t\t\t/>\n\t\t\t\t\t{/each}\n\t\t\t\t\t{#if isReadOnly}\n\t\t\t\t\t\t\n\t\t\t\t\t{/if}\n\t\t\t\t
\n\t\t\t{:else if pending}\n\t\t\t\t\n\t\t\t{:else}\n\t\t\t\t {\n\t\t\t\t\t\tonmessage?.(content);\n\t\t\t\t\t}}\n\t\t\t\t/>\n\t\t\t{/if}\n\t\t
\n\n\t\t\n\n\t\t\n\t\n\n\t*]:pointer-events-auto\"\n\t>\n\t\t{#if !draft.length && !messages.length && !sources.length && !loading && (currentModel.isRouter || (modelSupportsTools && $allBaseServersEnabled)) && activeExamples.length && !hideRouterExamples && !lastIsError && $mcpServersLoaded}\n\t\t\t\n\t\t\t\t{#each activeExamples as ex}\n\t\t\t\t\t startExample(ex)}>{ex.title}\n\t\t\t\t{/each}\n\t\t\t\n\t\t{/if}\n\t\t{#if shouldShowRouterFollowUps && !lastIsError}\n\t\t\t\n\t\t\t\t\n\t\t\t\t{#each routerFollowUps as followUp}\n\t\t\t\t\t startFollowUp(followUp)}\n\t\t\t\t\t>\n\t\t\t\t\t\t\n\t\t\t\t\t\t{followUp.title}\n\t\t\t\t{/each}\n\t\t\t\n\t\t{/if}\n\t\t{#if sources?.length && !loading}\n\t\t\t\n\t\t\t\t{#each sources as source, index}\n\t\t\t\t\t{#await source then src}\n\t\t\t\t\t\t {\n\t\t\t\t\t\t\t\tfiles = files.filter((_, i) => i !== index);\n\t\t\t\t\t\t\t}}\n\t\t\t\t\t\t/>\n\t\t\t\t\t{/await}\n\t\t\t\t{/each}\n\t\t\t\n\t\t{/if}\n\n\t\t
\n\t\t\t
\n\t\t\t\t{#if !loading && lastIsError}\n\t\t\t\t\t {\n\t\t\t\t\t\t\tif (lastMessage && lastMessage.ancestors) {\n\t\t\t\t\t\t\t\tonretry?.({\n\t\t\t\t\t\t\t\t\tid: lastMessage.id,\n\t\t\t\t\t\t\t\t});\n\t\t\t\t\t\t\t}\n\t\t\t\t\t\t}}\n\t\t\t\t\t/>\n\t\t\t\t{/if}\n\t\t\t
\n\t\t\t {\n\t\t\t\t\te.preventDefault();\n\t\t\t\t\thandleSubmit();\n\t\t\t\t}}\n\t\t\t\tclass={{\n\t\t\t\t\t\"relative flex w-full max-w-4xl flex-1 items-center rounded-xl border bg-gray-100 dark:border-gray-700 dark:bg-gray-800\": true,\n\t\t\t\t\t\"opacity-30\": isReadOnly,\n\t\t\t\t\t\"max-sm:mb-4\": focused && isVirtualKeyboard(),\n\t\t\t\t}}\n\t\t\t>\n\t\t\t\t{#if isRecording || isTranscribing}\n\t\t\t\t\t {\n\t\t\t\t\t\t\tisRecording = false;\n\t\t\t\t\t\t}}\n\t\t\t\t\t\tonconfirm={handleRecordingConfirm}\n\t\t\t\t\t\tonsend={handleRecordingSend}\n\t\t\t\t\t\tonerror={handleRecordingError}\n\t\t\t\t\t/>\n\t\t\t\t{:else if onDrag && isFileUploadEnabled}\n\t\t\t\t\t\n\t\t\t\t{:else}\n\t\t\t\t\t\n\t\t\t\t\t\t{#if lastIsError}\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t{:else}\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t{/if}\n\n\t\t\t\t\t\t{#if loading}\n\t\t\t\t\t\t\t {\n\t\t\t\t\t\t\t\t\thapticError();\n\t\t\t\t\t\t\t\t\tonstop?.();\n\t\t\t\t\t\t\t\t}}\n\t\t\t\t\t\t\t\tshowBorder={true}\n\t\t\t\t\t\t\t\tclassNames=\"absolute bottom-2 right-2 size-8 sm:size-7 self-end rounded-full border bg-white text-black shadow transition-none dark:border-transparent dark:bg-gray-600 dark:text-white\"\n\t\t\t\t\t\t\t/>\n\t\t\t\t\t\t{:else}\n\t\t\t\t\t\t\t{#if transcriptionEnabled}\n\t\t\t\t\t\t\t\t {\n\t\t\t\t\t\t\t\t\t\tisRecording = true;\n\t\t\t\t\t\t\t\t\t}}\n\t\t\t\t\t\t\t\t\taria-label=\"Start voice recording\"\n\t\t\t\t\t\t\t\t>\n\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t{/if}\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t{/if}\n\t\t\t\t\t
\n\t\t\t\t{/if}\n\t\t\t\n\t\t\t\n\t\t\t\t{#if models.find((m) => m.id === currentModel.id)}\n\t\t\t\t\t{#if loading && streamingToolCallName}\n\t\t\t\t\t\t\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t\tCalling tool\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t{availableTools.find((t) => t.name === streamingToolCallName)?.displayName ??\n\t\t\t\t\t\t\t\t\tstreamingToolCallName}\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t\n\t\t\t\t\t{:else if !currentModel.isRouter || !loading}\n\t\t\t\t\t\t {\n\t\t\t\t\t\t\t\tif (requireAuthUser()) {\n\t\t\t\t\t\t\t\t\te.preventDefault();\n\t\t\t\t\t\t\t\t}\n\t\t\t\t\t\t\t}}\n\t\t\t\t\t\t\tclass=\"inline-flex items-center gap-1 hover:underline\"\n\t\t\t\t\t\t>\n\t\t\t\t\t\t\t{#if currentModel.isRouter}\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t{currentModel.displayName}\n\t\t\t\t\t\t\t{:else}\n\t\t\t\t\t\t\t\tModel: {currentModel.displayName}\n\t\t\t\t\t\t\t\t{#if hasProviderOverride}\n\t\t\t\t\t\t\t\t\t{@const hubOrg =\n\t\t\t\t\t\t\t\t\t\tPROVIDERS_HUB_ORGS[providerOverride as keyof typeof PROVIDERS_HUB_ORGS]}\n\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\t\t{#if providerOverride === \"fastest\"}\n\t\t\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\t\t{:else if providerOverride === \"cheapest\"}\n\t\t\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\t\t{:else if hubOrg}\n\t\t\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\t\t{/if}\n\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t{/if}\n\t\t\t\t\t\t\t{/if}\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t\n\t\t\t\t\t{:else if showRouterDetails && streamingRouterMetadata?.route}\n\t\t\t\t\t\t\n\t\t\t\t\t\t\t\n\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t{streamingRouterMetadata.route}\n\t\t\t\t\t\t\t\n\n\t\t\t\t\t\t\twith\n\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t{streamingRouterModelName}\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t\n\t\t\t\t\t{:else}\n\t\t\t\t\t\t\n\t\t\t\t\t\t\t Routing\n\t\t\t\t\t\t\n\t\t\t\t\t{/if}\n\t\t\t\t{:else}\n\t\t\t\t\t\n\t\t\t\t\t\t{currentModel.id}\n\t\t\t\t\t\n\t\t\t\t{/if}\n\t\t\t\t{#if !messages.length && !loading}\n\t\t\t\t\tGenerated content may be inaccurate or false.\n\t\t\t\t{/if}\n\t\t\t\n\t\t\n\t\n\n\n\n" }, { "path": "src/lib/components/chat/FileDropzone.svelte", "content": "\n\n (onDragInner = true)}\n\tondragleave={() => (onDragInner = false)}\n\tondragover={(e) => {\n\t\te.preventDefault();\n\t}}\n\tclass=\"relative flex h-28 w-full max-w-4xl flex-col items-center justify-center gap-1 rounded-xl border-2 border-dotted {onDragInner\n\t\t? 'border-blue-200 !bg-blue-600/10 text-blue-600 *:pointer-events-none dark:border-blue-600 dark:bg-blue-600/20 dark:text-blue-600'\n\t\t: 'bg-gray-100 text-gray-500 dark:border-gray-500 dark:bg-gray-700 dark:text-gray-400'}\"\n>\n\t\n\t

Drop File to add to chat

\n\n" }, { "path": "src/lib/components/chat/ImageLightbox.svelte", "content": "\n\n\n\n\n\t\n\t\n\t\n\t\t\n\t\t\n\t\t\t\n\t\t\n\n\t\t\n\t\t\n\t\t e.stopPropagation()}\n\t\t/>\n\t\n\n" }, { "path": "src/lib/components/chat/MarkdownBlock.svelte", "content": "\n\n{#each renderedTokens as token}\n\t{#if token.type === \"text\"}\n\t\t\n\t\t{@html token.html}\n\t{:else if token.type === \"code\"}\n\t\t\n\t{/if}\n{/each}\n" }, { "path": "src/lib/components/chat/MarkdownRenderer.svelte", "content": "\n\n{#each blocks as block, index (loading && index === blocks.length - 1 ? `stream-${index}` : block.id)}\n\t\n{/each}\n" }, { "path": "src/lib/components/chat/MarkdownRenderer.svelte.test.ts", "content": "import MarkdownRenderer from \"./MarkdownRenderer.svelte\";\nimport { render } from \"vitest-browser-svelte\";\nimport { page } from \"@vitest/browser/context\";\n\nimport { describe, expect, it } from \"vitest\";\n\ndescribe(\"MarkdownRenderer\", () => {\n\tit(\"renders\", () => {\n\t\trender(MarkdownRenderer, { content: \"Hello, world!\" });\n\t\texpect(page.getByText(\"Hello, world!\")).toBeInTheDocument();\n\t});\n\tit(\"renders headings\", () => {\n\t\trender(MarkdownRenderer, { content: \"# Hello, world!\" });\n\t\texpect(page.getByRole(\"heading\", { level: 1 })).toBeInTheDocument();\n\t});\n\tit(\"renders links\", () => {\n\t\trender(MarkdownRenderer, { content: \"[Hello, world!](https://example.com)\" });\n\t\tconst link = page.getByRole(\"link\", { name: \"Hello, world!\" });\n\t\texpect(link).toBeInTheDocument();\n\t\texpect(link).toHaveAttribute(\"href\", \"https://example.com\");\n\t\texpect(link).toHaveAttribute(\"target\", \"_blank\");\n\t\texpect(link).toHaveAttribute(\"rel\", \"noreferrer\");\n\t});\n\tit(\"renders inline codespans\", () => {\n\t\trender(MarkdownRenderer, { content: \"`foobar`\" });\n\t\texpect(page.getByRole(\"code\")).toHaveTextContent(\"foobar\");\n\t});\n\tit(\"renders block codes\", () => {\n\t\trender(MarkdownRenderer, { content: \"```foobar```\" });\n\t\texpect(page.getByRole(\"code\")).toHaveTextContent(\"foobar\");\n\t});\n\tit(\"doesnt render raw html directly\", () => {\n\t\trender(MarkdownRenderer, { content: \"\" });\n\t\texpect(page.getByRole(\"button\").elements).toHaveLength(0);\n\t\t// htmlparser2 escapes disallowed tags\n\t\texpect(page.getByRole(\"paragraph\")).toHaveTextContent(\"\");\n\t});\n\tit(\"renders latex\", () => {\n\t\tconst { baseElement } = render(MarkdownRenderer, { content: \"$(oo)^2$\" });\n\t\texpect(baseElement.querySelectorAll(\".katex\")).toHaveLength(1);\n\t});\n\tit(\"does not render latex in code blocks\", () => {\n\t\tconst { baseElement } = render(MarkdownRenderer, { content: \"```\\n$(oo)^2$\\n```\" });\n\t\texpect(baseElement.querySelectorAll(\".katex\")).toHaveLength(0);\n\t});\n\tit(\"does not render latex in inline codes\", () => {\n\t\tconst { baseElement } = render(MarkdownRenderer, { content: \"`$oo` and `$bar`\" });\n\t\texpect(baseElement.querySelectorAll(\".katex\")).toHaveLength(0);\n\t});\n\tit(\"does not render latex across multiple lines\", () => {\n\t\tconst { baseElement } = render(MarkdownRenderer, { content: \"* $oo \\n* $aa\" });\n\t\texpect(baseElement.querySelectorAll(\".katex\")).toHaveLength(0);\n\t});\n\tit(\"renders latex with some < and > symbols\", () => {\n\t\tconst { baseElement } = render(MarkdownRenderer, { content: \"$foo < bar > baz$\" });\n\t\texpect(baseElement.querySelectorAll(\".katex\")).toHaveLength(1);\n\t});\n});\n" }, { "path": "src/lib/components/chat/MessageAvatar.svelte", "content": "\n\n\n\t\n\t\t\n\t\t\n\t\t\n\t\t\t\n\t\t\n\n\t\t\n\t\t\n\t\t\t\n\t\t\t\n\t\t\t\t\n\t\t\t\t\n\t\t\t\n\t\t\n\t\n\n\t\n\t\t\n\t\t\n\t\t\t\n\t\t\t\n\t\t\t\n\t\t\n\t\n\n" }, { "path": "src/lib/components/chat/ModelSwitch.svelte", "content": "\n\n\n\t\n\t\tThis model is no longer available. Switch to a new one to continue this conversation:\n\t\n\t
\n\t\t\n\t\t\t{#each models as model}\n\t\t\t\t\n\t\t\t{/each}\n\t\t\n\t\t\n\t\t\tAccept\n\t\t\n\t
\n\n" }, { "path": "src/lib/components/chat/OpenReasoningResults.svelte", "content": "\n\n{#snippet icon()}\n\t\n\t\t\n\t\n{/snippet}\n\n\n\t\n\t (isOpen = !isOpen)}\n\t>\n\t\t{#if isOpen}\n\t\t\t\n\t\t\t\n\t\t\t\t\n\t\t\t\n\t\t{:else}\n\t\t\t\n\t\t\t\n\t\t\t\t{content\n\t\t\t\t\t.replace(/[#*`~[\\]]/g, \"\")\n\t\t\t\t\t.replace(/\\n+/g, \" \")\n\t\t\t\t\t.trim()}\n\t\t\t\n\t\t{/if}\n\t\n\n" }, { "path": "src/lib/components/chat/ToolUpdate.svelte", "content": "\n\n{#snippet icon()}\n\t{#if toolSuccess}\n\t\t\n\t{:else}\n\t\t\n\t{/if}\n{/snippet}\n\n{#if toolFnName}\n\t\n\t\t\n\t\t
\n\t\t\t (isOpen = !isOpen)}\n\t\t\t>\n\t\t\t\t\n\t\t\t\t\t{toolError ? \"Error calling\" : toolDone ? \"Called\" : \"Calling\"} tool\n\t\t\t\t\t\n\t\t\t\t\t\t{availableTools.find((entry) => entry.name === toolFnName)?.displayName ?? toolFnName}\n\t\t\t\t\t\n\t\t\t\t\n\t\t\t\t{#if isExecuting && toolProgress}\n\t\t\t\t\t{progressLabel}\n\t\t\t\t{/if}\n\t\t\t\n\n\t\t\t (isOpen = !isOpen)}\n\t\t\t\taria-label={isOpen ? \"Collapse\" : \"Expand\"}\n\t\t\t>\n\t\t\t\t\n\t\t\t\n\t\t
\n\n\t\t\n\t\t{#if isOpen}\n\t\t\t
\n\t\t\t\t{#each tool as update, i (`${update.subtype}-${i}`)}\n\t\t\t\t\t{#if update.subtype === MessageToolUpdateType.Call}\n\t\t\t\t\t\t
\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\tInput\n\t\t\t\t\t\t\t
\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t
{formatValue(\n\t\t\t\t\t\t\t\t\t\tupdate.call.parameters\n\t\t\t\t\t\t\t\t\t)}
\n\t\t\t\t\t\t\t
\n\t\t\t\t\t\t\n\t\t\t\t\t{:else if update.subtype === MessageToolUpdateType.Error}\n\t\t\t\t\t\t
\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\tError\n\t\t\t\t\t\t\t
\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t
{update.message}
\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t\n\t\t\t\t\t{:else if isMessageToolResultUpdate(update) && update.result.status === ToolResultStatus.Success && update.result.display}\n\t\t\t\t\t\t
\n\t\t\t\t\t\t\t
\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\tOutput\n\t\t\t\t\t\t\t\t
\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t
\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t{#each parseToolOutputs(update.result.outputs) as parsedOutput}\n\t\t\t\t\t\t\t\t\t
\n\t\t\t\t\t\t\t\t\t\t{#if parsedOutput.text}\n\t\t\t\t\t\t\t\t\t\t\t{parsedOutput.text}\n\t\t\t\t\t\t\t\t\t\t{/if}\n\n\t\t\t\t\t\t\t\t\t\t{#if parsedOutput.images.length > 0}\n\t\t\t\t\t\t\t\t\t\t\t
\n\t\t\t\t\t\t\t\t\t\t\t\t{#each parsedOutput.images as image, imageIndex}\n\t\t\t\t\t\t\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t\t\t\t\t{/each}\n\t\t\t\t\t\t\t\t\t\t\t
\n\t\t\t\t\t\t\t\t\t\t{/if}\n\n\t\t\t\t\t\t\t\t\t\t{#if parsedOutput.metadata.length > 0}\n\t\t\t\t\t\t\t\t\t\t\t
{formatValue(\n\t\t\t\t\t\t\t\t\t\t\t\t\tObject.fromEntries(parsedOutput.metadata)\n\t\t\t\t\t\t\t\t\t\t\t\t)}
\n\t\t\t\t\t\t\t\t\t\t{/if}\n\t\t\t\t\t\t\t\t\t
\n\t\t\t\t\t\t\t\t{/each}\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t\n\t\t\t\t\t{:else if isMessageToolResultUpdate(update) && update.result.status === ToolResultStatus.Error && update.result.display}\n\t\t\t\t\t\t
\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\tError\n\t\t\t\t\t\t\t
\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t\t\t
{update.result\n\t\t\t\t\t\t\t\t\t\t.message}
\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t\n\t\t\t\t\t{/if}\n\t\t\t\t{/each}\n\t\t\t\n\t\t{/if}\n\t\n{/if}\n" }, { "path": "src/lib/components/chat/UploadedFile.svelte", "content": "\n\n{#if showModal && isClickable}\n\t\n\t (showModal = false)}>\n\t\t{#if isImage(file.mime)}\n\t\t\t{#if file.type === \"hash\"}\n\t\t\t\t\n\t\t\t{:else}\n\t\t\t\t\n\t\t\t\t\n\t\t\t{/if}\n\t\t{:else if isPlainText(file.mime)}\n\t\t\t
\n\t\t\t\t
\n\t\t\t\t\t\n\t\t\t\t\t

{file.name}

\n\t\t\t\t
\n\t\t\t\t{#if file.mime === \"application/vnd.chatui.clipboard\"}\n\t\t\t\t\t

\n\t\t\t\t\t\tIf you prefer to inject clipboard content directly in the chat, you can disable this\n\t\t\t\t\t\tfeature in the\n\t\t\t\t\t\tsettings page.\n\t\t\t\t\t

\n\t\t\t\t{/if}\n\t\t\t\t (showModal = false)}\n\t\t\t\t>\n\t\t\t\t\t\n\t\t\t\t\n\t\t\t\t{#if file.type === \"hash\"}\n\t\t\t\t\t{#await fetch(urlNotTrailing + \"/output/\" + file.value).then((res) => res.text())}\n\t\t\t\t\t\t
\n\t\t\t\t\t\t\t\n\t\t\t\t\t\t
\n\t\t\t\t\t{:then result}\n\t\t\t\t\t\t{result}\n\t\t\t\t\t{/await}\n\t\t\t\t{:else}\n\t\t\t\t\t{atob(file.value)}\n\t\t\t\t{/if}\n\t\t\t
\n\t\t{/if}\n\t\n{/if}\n\n isClickable && (showModal = true)}\n\tonkeydown={(e) => {\n\t\tif (!isClickable) {\n\t\t\treturn;\n\t\t}\n\t\tif (e.key === \"Enter\" || e.key === \" \") {\n\t\t\tshowModal = true;\n\t\t}\n\t}}\n\tclass:clickable={isClickable}\n\trole=\"button\"\n\ttabindex=\"0\"\n>\n\t
\n\t\t{#if isImage(file.mime)}\n\t\t\t
\n\t\t\t\t\n\t\t\t
\n\t\t{:else if isAudio(file.mime)}\n\t\t\t\n\t\t{:else if isVideo(file.mime)}\n\t\t\t\n\t\t\t\t\n\t\t\t\t\n\t\t\t
\n\t\t{:else if isPlainText(file.mime)}\n\t\t\t\n\t\t\t\t\n\t\t\t\t\t\n\t\t\t\t\n\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t\t\t\t{truncateMiddle(file.name, 28)}\n\t\t\t\t\t
\n\t\t\t\t\t{#if file.mime === \"application/vnd.chatui.clipboard\"}\n\t\t\t\t\t\t
Clipboard source
\n\t\t\t\t\t{:else}\n\t\t\t\t\t\t
{file.mime}
\n\t\t\t\t\t{/if}\n\t\t\t\t
\n\t\t\t\n\t\t{:else if file.mime === \"application/octet-stream\"}\n\t\t\t\n\t\t\t\t\n\t\t\t\t\t\n\t\t\t\t\n\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t\t\t\t{truncateMiddle(file.name, 28)}\n\t\t\t\t\t
\n\t\t\t\t\t
File type could not be determined
\n\t\t\t\t
\n\t\t\t\t\n\t\t\t\t\t\n\t\t\t\t\n\t\t\t\n\t\t{:else}\n\t\t\t\n\t\t\t\t\n\t\t\t\t\t\n\t\t\t\t\n\t\t\t\t
\n\t\t\t\t\t
\n\t\t\t\t\t\t{truncateMiddle(file.name, 28)}\n\t\t\t\t\t
\n\t\t\t\t\t
{file.mime}
\n\t\t\t\t
\n\t\t\t\n\t\t{/if}\n\t\t\n\t\t{#if canClose}\n\t\t\t {\n\t\t\t\t\te.preventDefault();\n\t\t\t\t\te.stopPropagation();\n\t\t\t\t\tonclose?.();\n\t\t\t\t}}\n\t\t\t>\n\t\t\t\t\n\t\t\t\n\t\t{/if}\n\t\n\n" }, { "path": "src/lib/components/chat/UrlFetchModal.svelte", "content": "\n\n{#if open}\n\t\n\t\t{#snippet children()}\n\t\t\t {\n\t\t\t\t\te.preventDefault();\n\t\t\t\t\thandleSubmit();\n\t\t\t\t}}\n\t\t\t>\n\t\t\t\t
\n\t\t\t\t\t

Add from URL

\n\t\t\t\t\t\n\t\t\t\t
\n\n\t\t\t\t
\n\t\t\t\t\t
\n\n\t\t\t\t{#if errorMsg}\n\t\t\t\t\t

{errorMsg}

\n\t\t\t\t{/if}\n\t\t\t\t

Only HTTPS. Max 10MB.

\n\n\t\t\t\t
\n\t\t\t\t\t\n\t\t\t\t\t\tCancel\n\t\t\t\t\t\n\t\t\t\t\t\n\t\t\t\t\t\t{#if loading}Fetching…{:else}Add{/if}\n\t\t\t\t\t\n\t\t\t\t
\n\t\t\t\n\t\t{/snippet}\n\t\n{/if}\n\n\n" }, { "path": "src/lib/components/chat/VoiceRecorder.svelte", "content": "\n\n
\n\t\n\t\n\t\t\n\t\n\n\t\n\t
\n\t\t{#if isTranscribing}\n\t\t\t
\n\t\t\t\t\n\t\t\t
\n\t\t{:else}\n\t\t\t\n\t\t{/if}\n\t
\n\n\t\n\t\n\t\t{#if isTranscribing}\n\t\t\t\n\t\t{:else if isTouchDevice}\n\t\t\t\n\t\t{:else}\n\t\t\t\n\t\t{/if}\n\t\n
\n" }, { "path": "src/lib/components/icons/IconBurger.svelte", "content": "\n\n\n\n" }, { "path": "src/lib/components/icons/IconCheap.svelte", "content": "\n\n\n\t\n\n" }, { "path": "src/lib/components/icons/IconChevron.svelte", "content": "\n\n\n\t\n\n" }, { "path": "src/lib/components/icons/IconDazzled.svelte", "content": "\n\n\n\t\n\t\n\t\n\t\t\n\t\t\t\n\t\t\t\n\t\t\n\t\n\n" }, { "path": "src/lib/components/icons/IconFast.svelte", "content": "\n\n\n\t\n\n" }, { "path": "src/lib/components/icons/IconLoading.svelte", "content": "\n\n
\n\t
\n\t\n\t\n\n" }, { "path": "src/lib/components/icons/IconMCP.svelte", "content": "\n\n\n\t\n\t\t\n\t\t\n\t\n\n" }, { "path": "src/lib/components/icons/IconMoon.svelte", "content": "\n\n\n\t\n\n" }, { "path": "src/lib/components/icons/IconNew.svelte", "content": "\n\n\n" }, { "path": "src/lib/components/icons/IconOmni.svelte", "content": "\n\n