Full Code of global-agent-hackathon/global-agent-hackathon-may-2025 for AI

Repository: global-agent-hackathon/global-agent-hackathon-may-2025
Branch: main
Commit: 1d6037323635
Files: 13
Total size: 21.6 KB

Directory structure:
gitextract_4j08bamc/

├── .github/
│   ├── ISSUE_TEMPLATE/
│   │   └── submission-template.md
│   └── workflows/
│       └── auto-close.yml
├── README.md
├── resources/
│   ├── agno.md
│   ├── browser-use.md
│   ├── firecrawl.md
│   ├── graphlit.md
│   ├── lutra-ai.md
│   ├── mem0.md
│   ├── potpie.md
│   └── unwindai.md
├── submissions/
│   └── .gitkeep
└── winners/
    └── .gitkeep

================================================
FILE CONTENTS
================================================

================================================
FILE: .github/ISSUE_TEMPLATE/submission-template.md
================================================
---
name: Hackathon Project Submission
about: Submit your project for the Global Agent Hackathon
title: "[Project Submission] Your Project Title"
labels: submission
assignees: ''

---

# Project Title
A clear, concise, and engaging title for your project.

## Overview of the Idea
Briefly describe your project in simple terms. What problem or opportunity does it address?

## Project Goal
What do you hope to achieve or demonstrate with this project?

## How It Works
Provide a step-by-step explanation or high-level breakdown of your Agentic System and how users will interact with it:

- **User Flow**: Clearly describe the experience from the user's perspective
- **Core Functionality**: List the key features your agent will offer
- **Multimodal Elements** (if applicable): Specify if you're using text, images, audio, video, or combinations

## Tools Used
List what you're planning to use, including any new tools you need to create.

## UI Approach
Describe the user interface for your project.

## Visuals
Add mockups, diagrams, or sketches that demonstrate your idea or flow.

## Team Information
- **Team Lead**: GitHub handle and role
- **Team Members**: GitHub handles and roles
- **Background/Experience**: Brief team background

## Prize Category (leave blank, to be assigned by judges)
- [ ] Best use of Agno
- [ ] Best use of Firecrawl
- [ ] Best use of Mem0
- [ ] Best use of Graphlit
- [ ] Best use of Browser Use
- [ ] Best use of Potpie
- [ ] Best Overall Project

## Demo Video Link
Once completed, add a link to your 2-3 minute demo video.

## Additional Notes
Any other information you'd like to share about your project.


================================================
FILE: .github/workflows/auto-close.yml
================================================
name: Auto-close PRs and Issues

on:
  issues:
    types: [opened]
  pull_request:
    types: [opened]

permissions:
  issues: write
  pull-requests: write

jobs:
  auto-close:
    runs-on: ubuntu-latest
    steps:
      - name: Close new PRs/issues
        uses: actions/github-script@v7
        with:
          script: |
            const isPR = !!context.payload.pull_request;
            const number = isPR ? context.payload.pull_request.number : context.payload.issue.number;

            // Close the issue or PR
            if (isPR) {
              await github.rest.pulls.update({
                ...context.repo,
                pull_number: number,
                state: 'closed'
              });
            } else {
              await github.rest.issues.update({
                ...context.repo,
                issue_number: number,
                state: 'closed'
              });
            }

            await github.rest.issues.createComment({
              ...context.repo,
              issue_number: number,
              body: `Thank you for your contribution. The last day to participate in the hackathon was 1st June 2025. We are no longer accepting new issues or pull requests. We will be announcing the winners soon.`
            });


================================================
FILE: README.md
================================================
<div align="center" id="top">
  <img src="./.assets/may-2025-banner.svg" alt="Agno" width="750" height="auto">
</div>

# The Global Agent Hackathon

Welcome to the May 2025 edition of the Global Agent Hackathon!

Join us for a month-long, open-source AI Agent Hackathon — open to all builders and dreamers working on agents, RAG, tool use, and multi-agent systems.

Powered by [Agno](https://agno.link/gh), [Exa](https://exa.ai), [Firecrawl](https://www.firecrawl.dev/), [Mem0](http://mem0.ai), [Unwind AI](https://www.theunwindai.com), [Browser Use](https://browser-use.com), [mcp.run](https://www.mcp.run), [Graphlit](https://www.graphlit.com/), [Lutra AI](https://www.lutra.ai), [Potpie](https://potpie.ai) and more.

## 💰 Build something cool, win over $25,000 in cash and credits

Win over $25,000 in cash and credits for your favorite AI tools for the most ambitious Agent projects!

## Cash Prizes

### Best Use of Agno ([see details](/resources/agno.md)):

Projects showcasing the best use case of Agno will get following cash prizes:
- 🏆 **GRAND PRIZE**: $5,000 🏆
- 🥇 1 winner: $2,000
- 🥈 5 winners: $1,000 each
- 🥉 10 winners: $500 each
- 🏅 10 winners: $300 each

The $300, $500 and $1,000 prizes will be awarded on a rolling basis. The $2,000 and $5,000 prizes will be awarded at the end of the hackathon.

Judges: Dirk Brand, Yash Solanky, Monali Dambre

### Best use of Browser Use ([see details](/resources/browser-use.md)):

Projects showcasing the best use case of Browser Use will get following cash prizes:
- 🏆 **GRAND PRIZE**: $2,500 🏆
- 🥉 2 winners: $500 each
- 🥈 10 winners: $150 each

The $150 prizes will be awarded on a rolling basis. The $500 and $2,500 prizes will be awarded at the end of the hackathon.

Judges: Larsen Cundric, Mert Ünsal and Aitor Mato

### Best use of Potpie 🥧 ([see details](/resources/potpie.md)):

Projects showcasing the best use case of Potpie will get following cash prizes:
- 🏆 **GRAND PRIZE**: $200 🏆
- 🥉 3 winners: $100 each
- 🥈 5 winners: 3 months of Potpie Pro FREE!

All the prizes will be awarded at the end of the hackathon.

Judges: Aditi

## Credits

### Browserbase - $1200 value:

- Startup plan for free for a whole year to the best project built using Browserbase

Judges: Lindsay Gilson

### Exa - Free $50 credits for everyone:

Every participant gets free $50 Exa API credits!

- To get the credits use the coupon code `EXA50API` by heading to the [Exa API Dashboard](https://dashboard.exa.ai/billing).

### Best use of Firecrawl:

- Every one is entitled to 5k Firecrawl credits, enter the coupon code `hackathon5k` on the [Firecrawl Dashboard](https://www.firecrawl.dev/app/settings)

### Best use of Mem0:

- Mem0 is offering pro plan for free for 3 months to the best project built using Mem0. ([details coming soon](/resources/mem0.md))

Judges: Taranjeet Singh

### MCP Run:

- [MCP.Run](https://www.mcp.run) is offering a free 3 month pro account to any participant, using `AGNOHACK2025` at checkout. Thank you MCP.Run for sponsoring the hackathon!

### Best use of Graphlit:

- Please [see details](/resources/graphlit.md)

### Best use of Lutra AI:

- [Details coming soon](/resources/lutra-ai.md)

## Groq Credits

Groq is the lowest cost inference provider with no tradeoffs on speed and they're offering $10 in Groq API credits to any participant (millions of tokens for all available models, including Llama 4 Scout and Maverick) and $300 in Groq API credits to the best project built using Groq.

- [Fill out this form for $10 in Groq API credits](https://forms.gle/1pBpXsi2j8btk1Rd6)
- Best use of Groq will get $300 in Groq API credits + feature on [Groq Applications Showcase](https://console.groq.com/docs/showcase-applications)

## Special Recognitions

- The **top 5 projects** will be featured in the trending [Awesome LLM Apps repo](https://github.com/Shubhamsaboo/awesome-llm-apps), giving your work massive visibility in the AI community.

Judges: Shubham Saboo

## 📚 Resources

Resources for each sponsor are available in the [resources](/resources) folder.

## ⏱️ Timeline

- **April 15th** - Hackathon starts
- **June 1st** - Hackathon ends

## 🚀 How to Participate

1. **Submit your project** by creating a GitHub issue in this repo following the provided [template](./.github/ISSUE_TEMPLATE/submission-template.md).
2. **Get category assignment**
    - The Hackathon team will review your submission and assign it to a prize category.
    - You can update your issue as you build, and we'll update the prize category accordingly.
    - Prize categories can be updated until the hackathon ends.
    - Categories include Best use of Agno, Firecrawl, Mem0, Graphlit and Browser Use.
3. **Build your project** and submit a PR for your project in the submissions folder - we'll review, judge, and provide feedback. Note: you can start building as we review your submission.
4. **When your PR is approved**, you join the pool to win the assigned cash prize and/or credits.
    - Checkout the [How to Win](#how-to-win) section for more details.
5. **Winning projects** will be moved to the **winners** folder.

## 🏆 How to Win

Your submission will be judged on:

- **Completed PR** with working code in the Hackathon GitHub repo.
    - A **judge** will review the PR and provide feedback.
    - Once approved, the PR will be merged and the project will be eligible for the prize pool.
- **Demo Video** (2–3 mins) showing your agent in action.
- **Comprehensive README.md** including:
  - Overview of the idea
  - Explanation of how it works
  - Technologies/tools used
  - All setup instructions (API keys, signups, etc.)
- **Code Quality**: Error-free execution and passing basic review.
- **Completeness**: Core functionality implemented as described in your original submission.

Feel free to share your progress on X or LinkedIn with the hashtag #GlobalAgentHackathon — even small updates can get you noticed and help build community momentum. This is completely optional, but always welcome and appreciated!

## 📌 Submission Criteria

Include the following in your submission:

**Project Title:** A clear, concise, and engaging title for your project.

**Overview of the Idea:** Briefly describe your project in simple terms. What problem or opportunity does it address?

**Project Goal:** What do you hope to achieve or demonstrate with this project?

**How It Works:** Provide a step-by-step explanation or high-level breakdown of your Agentic System and how users will interact with it:
- **User Flow**: Clearly describe the experience from the user's perspective
- **Core Functionality**: List the key features your agent will offer
- **Multimodal Elements** (if applicable): Specify if you're using text, images, audio, video, or combinations

**Tools Used:** List what you're planning to use, including any new tools you need to create.

**UI Approach:** Describe the user interface for your project.

**Visuals:** Add mockups, diagrams, or sketches that demonstrate your idea or flow.

**Team Information:**
- **Team Lead:** GitHub handle and role
- **Team Members:** GitHub handles and roles
- **Background/Experience:** Brief team background

## 🤝 Want to Partner?

If you're an AI company and want to help shape the next generation of Agent builders - we'd love to work with you!

Reach out to support@agno.com to get involved.

Let's make this the largest open-source AI Agent hackathon ever!

**See you in the arena.**


================================================
FILE: resources/agno.md
================================================
<div align="center" id="top">
  <a href="https://docs.agno.com">
    <picture>
      <source media="(prefers-color-scheme: dark)" srcset="https://agno-public.s3.us-east-1.amazonaws.com/assets/logo-dark.svg">
      <source media="(prefers-color-scheme: light)" srcset="https://agno-public.s3.us-east-1.amazonaws.com/assets/logo-light.svg">
      <img src="https://agno-public.s3.us-east-1.amazonaws.com/assets/logo-light.svg" alt="Agno">
    </picture>
  </a>
</div>
<div align="center">
  <a href="https://docs.agno.com">📚 Documentation</a> &nbsp;|&nbsp;
  <a href="https://docs.agno.com/examples/introduction">💡 Examples</a> &nbsp;|&nbsp;
  <a href="https://github.com/agno-agi/agno/stargazers">🌟 Star Us</a>
</div>

<br />

[Agno](https://docs.agno.com) is a lightweight library for building Agents with memory, knowledge, tools and reasoning.

## 💰 Build something cool with Agno to win up to $20,000 in cash

Projects showcasing the best use of Agno will be eligible for the following cash prizes:

- 🏅 10 winners: $300 each
- 🥉 10 winners: $500 each
- 🥈 5 winners: $1,000 each
- 🥇 1 winner: $2,000
- 🏆 **GRAND PRIZE**: $5,000 🏆

The $300, $500 and $1,000 prizes will be awarded on a rolling basis until the prize pool is exhausted. The $2,000 and $5,000 prizes will be awarded at the end of the hackathon.

## Building Agents with Agno

If you're new to Agno, start by building your [first Agent](https://docs.agno.com/introduction/agents), then chat with it on the [playground](https://docs.agno.com/introduction/playground) and monitor it on [agno.com](https://docs.agno.com/introduction/monitoring).

## Cursor Setup

When building Agno agents, using Agno documentation as a source in Cursor is a great way to speed up your development.

1. In Cursor, go to the settings or preferences section.
2. Find the section to manage documentation sources.
3. Add `https://docs.agno.com` to the list of documentation URLs.
4. Save the changes.

Now, Cursor will have access to the Agno documentation.

## Documentation, Community & More examples

- <a href="https://docs.agno.com" target="_blank" rel="noopener noreferrer">Documentation</a>
- <a href="https://github.com/agno-agi/agno/tree/main/cookbook" target="_blank" rel="noopener noreferrer">Cookbook with 100+ examples</a>
- <a href="https://community.agno.com/" target="_blank" rel="noopener noreferrer">Community forum</a>
- <a href="https://discord.gg/4MtYHHrgA8" target="_blank" rel="noopener noreferrer">Discord</a>


================================================
FILE: resources/browser-use.md
================================================


================================================
FILE: resources/firecrawl.md
================================================


================================================
FILE: resources/graphlit.md
================================================
# Best Use of Graphlit

We're excited to sponsor the Global Agent Hackathon with Agno.

We will be offering credits on your paid Graphlit project for the best use of Graphlit in Agno agents.

You can use [Agno](https://docs.agno.com/tools/mcp) agents with MCP servers like Graphlit [Model Context Protocol (MCP) Server](https://github.com/graphlit/graphlit-mcp-server)

Ingest anything from Slack, Discord, websites, Google Drive, email, Jira, Linear or GitHub into a Graphlit project - and then search and retrieve relevant knowledge within your Agno agent.

Documents (PDF, DOCX, PPTX, etc.) and HTML web pages will be extracted to Markdown upon ingestion. Audio and video files will be transcribed upon ingestion.

Web crawling and web search are built-in as MCP tools, with no need to integrate other tools like Firecrawl, Exa, etc. separately.

Here's a [YouTube video](https://www.youtube.com/watch?v=Or-QqonvcAs&t=4s) showing the capabilities of the Graphlit MCP Server with the Goose MCP client, which may give you ideas for your own AI agents.

For any questions on using the Graphlit MCP Server, please join our [Discord](https://discord.gg/ygFmfjy3Qx) community and post on the #mcp channel.

Let us know if you need any other data connectors, or have ideas on new MCP tools that would make your Agno agent even more useful.

## Tools

### Retrieval

- Query Contents
- Query Collections
- Query Feeds
- Query Conversations
- Retrieve Relevant Sources
- Retrieve Similar Images
- Visually Describe Image

### RAG

- Prompt LLM Conversation

### Extraction

- Extract Structured JSON from Text

### Ingestion

- Files
- Web Pages
- Messages
- Posts
- Emails
- Issues
- Text
- Memory (Short-Term)

### Data Connectors

- Microsoft Outlook email
- Google Mail
- Notion
- Reddit
- Linear
- Jira
- GitHub Issues
- Google Drive
- OneDrive
- SharePoint
- Dropbox
- Box
- GitHub
- Slack
- Microsoft Teams
- Discord
- Twitter/X
- Podcasts (RSS)

### Web

- Web Crawling
- Web Search (including Podcast Search)
- Web Mapping
- Screenshot Page

### Notifications

- Slack
- Email
- Webhook
- Twitter/X


================================================
FILE: resources/lutra-ai.md
================================================


================================================
FILE: resources/mem0.md
================================================


================================================
FILE: resources/potpie.md
================================================
# Best Use of Potpie 🥧

We’re thrilled to be part of the Global Agent Hackathon with Agno! 🧠⚙️

We're offering the following prizes for the best use of [Potpie](https://github.com/potpie-ai/potpie) during the hackathon.

Projects showcasing the best use case of Potpie will get following cash prizes:
- 🏆 **GRAND PRIZE**: $200 🏆
- 🥉 3 winners: $100 each
- 🥈 5 winners: 3 months of Potpie Pro FREE

All the prizes will be awarded at the end of the hackathon.

How well you market your solution will also be judged in order to win the above prize.

Potpie is an **open-source agent platform** built specifically for developers and their codebases. It turns a single prompt into a powerful, context-aware agent that understands your entire repo.

You can use Potpie to build agents that:
- Understand and navigate your codebase with a knowledge graph
- Perform debugging, system design, test generation, and more
- Connect with tools like GitHub, Jira, and Linear to create dev workflows

If you’ve ever wanted a dev tool that *actually* understands your stack — Potpie is it.

---

## How Potpie Works

Potpie indexes your codebase and builds a semantic **knowledge graph**, allowing agents to:
- Traverse and reason across code
- Link functions, classes, APIs, components, and files
- Provide rich context-aware answers and workflows

Potpie integrates tightly with:
- GitHub (repos, PRs, commits)
- Linear (issues, tickets)

Use Potpie with Agno’s MCP to inject deep codebase intelligence into your agents.

---

## Building Custom Agents

### Potpie Tools

You can create custom agents from a single prompt on app.potpie.ai and those agents select the tools on their own, but sometimes we need to include or remove certain tools while iteration on the agent outputs, prompts might need to be updated to include specific guidance for their execution plan etc  For those times, here are the tools that Potpie agents use internally and a brief description about their usage. 

- **get_code_from_probable_node_name**: Fetches code based on a likely node name match in the knowledge graph.
- **get_code_from_multiple_node_ids**: Returns code snippets for a list of node IDs.
- **ask_knowledge_graph_queries**: Enables natural language querying over the knowledge graph.
- **get_nodes_from_tags**: Lists nodes associated with specific tags like `API`, `Database`, etc.
- **get_code_graph_from_node_id**: Generates a local dependency graph starting from a specific node ID.
- **change_detection**: Analyzes code changes and summarizes diffs with metadata.
- **get_code_file_structure**: Returns the directory and file structure of the ingested codebase.
- **get_node_neighbours_from_node_id**: Retrieves neighboring nodes in the knowledge graph from a given node ID.
- **get_linear_issue**: Fetches details of a Linear issue by ID. (requires linear API key to be set in Key Management screen) 
- **update_linear_issue**: Updates status, priority, or description of a Linear issue.(requires linear API key to be set in Key Management screen) 
- **intelligent_code_graph**: Generates a filtered code graph, good for reducing context in larger codebases. 
- **think**: A reflective reasoning tool for maintaining context or brainstorming within an agent's workflow.
- **github_create_branch**: Creates a new GitHub branch with metadata context.
- **github_update_branch**: Pushes code updates to an existing GitHub branch.
- **github_create_pull_request**: Opens a GitHub PR with generated title, description, and branch metadata.
- **github_add_pr_comments**: Adds inline review comments to a GitHub PR based on context or code diffs.


### Example Dev Workflows 

- Generate unit/integration tests
- Suggest design improvements
- Summarize PRs
- Auto-create GitHub issues or Linear tickets based on Potpie agent results
- Review workflows: Extract change impact and recommend fixes
- Automate integrations that follow the same patterns (E.g. Adding a memory provider into an Agent framework etc) 

### Ingestion

- GitHub Repos Issues & Pull Requests
- Ticketing Systems (Linear)
- Web pages (Direct link) 

---
## Generating an API Key

To generate an API key, follow these steps:

1. Navigate to [app.potpie.ai](https://app.potpie.ai).
2. Click on your username in the bottom left corner.
3. Select **Key Management** from the menu.
4. In the "API Key Management" section, click on **Generate API Key**.

Your API key will be created and displayed for use.

## Using the API

Postman collection is hosted [here](https://drive.google.com/file/d/1DMF2A5oy8zeNb0xQMUn82fcF2CVvc__L/view?usp=sharing)
Video walkthrough [here](https://x.com/runtimeHorror/status/1897144432404914215)

To use the API, follow these steps:

1. **Base URL**: Use `https://production-api.potpie.ai/` as the base URL.

2. **Parse API**:
   - **API Path**: `/api/v2/parse`
   - **Request Method**: POST
   - **Request Body**:
     ```json
     {
       "repo_name": "potpie-ai/potpie",
       "branch_name": "main"
     }
     ```
   - The Parse API will return a **project ID**.

3. **Get Parsing Status**:
   - **API Path**: `/api/v2/parsing-status/{project_id}`
   - **Request Method**: GET
   - **Query Parameters**:
     - `project_id`: The ID returned from the Parse API.
     - `x-api-key`: Your API key (in the header).
   - Wait until the project status is "ready".


### List of Agents

You can choose from the following agents:

- **Codebase Q&A Agent**: `id="codebase_qna_agent"`
- **Debugging with Knowledge Graph Agent**: `id="debugging_agent"`
- **Unit Test Agent**: `id="unit_test_agent"`
- **Integration Test Agent**: `id="integration_test_agent"`
- **Low-Level Design Agent**: `id="LLD_agent"`
- **Code Changes Agent**: `id="code_changes_agent"`
- **Code Generation Agent**: `id="code_generation_agent"`

- For custom agents created on the Potpie dashboard, you can simply copy their Agent id using the copy button on the Custom Agent screen. 

4. **Send Message**:

   - **API Path**: `/api/v2/project/{{project_id}}/message`
   - **Request Method**: POST
   - **Request Body**:
     ```json
     {
       "content": "your_message_content",
       "node_ids": [] //leave this empty
     }
     ```
   - **Query Parameters**:
     - `conversation_id`: The ID of the conversation created in the previous step.
     - `x-api-key`: Your API key (in the header).

Note: 
* Try to use the /parse API sparingly, every invocation of this API will trigger the processing step if the knowledge graph is not upto date.
* We recommend that once you have a project in the 'ready' state, you hardcode the project id in this API, this avoids the added latency of parsing the codebase everytime unnecessarily. 


---

## 🎥 Need Inspo?

Here’s a [video walkthrough](https://youtu.be/a4rInO0xa54?feature=shared) of how Potpie can automate integration work in repos like Langflow. 

---

## 💬 Questions?

Join our [Discord](https://discord.com/invite/ryk5CMD5v6) — we’re happy to help you build custom agent magic.


================================================
FILE: resources/unwindai.md
================================================


================================================
FILE: submissions/.gitkeep
================================================


================================================
FILE: winners/.gitkeep
================================================