Team Ai
Apppublic

Agents-MCP-Hackathon/OutageOdyssey

sourceHugging Faceapache-2.0updated 1y agoView on Hugging Face
6likes
App README

tags:

  • —mcp-server-track
  • —agent-demo-track
  • —custom-component-track ---

🚨 Outage Odyssey 🧩

Transform boring IT incidents into hilarious learning experiences!

Outage Odyssey is an innovative hackathon project that converts mundane IT incident reports into engaging, educational comic stories. By leveraging AI agents and visual generation, it transforms technical troubleshooting into an entertaining and memorable learning experience.

🎯 What it Does

Outage Odyssey takes your IT incidents through a magical transformation:

  1. 1.🔍 Analyze - Deep dive into your incident details
  2. 2.🧩 Find - Identify the root cause with precision
  3. 3.🛠️ Solve - Generate actionable solution recommendations
  4. 4.📚 Create - Craft entertaining comic narratives
  5. 5.🎨 Generate - Produce comic-style visual panels
  6. 6.🎬 Deliver - Present as an educational comic strip

✨ Key Features

  • —Smart Incident Analysis - Automatically categorizes and analyzes IT incidents
  • —Root Cause Detection - Uses AI to identify underlying problems
  • —Solution Recommendations - Generates actionable step-by-step resolution guides
  • —Comic Story Generation - Transforms technical details into engaging narratives
  • —Visual Comic Panels - Generates comic-style illustrations
  • —Multi-format Support - Handles text, PDFs, logs, and documentation
  • —Interactive UI - Beautiful Gradio interface with incident templates
  • —Educational Focus - Makes learning from failures fun and memorable

Demo Video Link

Here is the demo video link: https://www.loom.com/share/93182321fd5b40afa5d20696e61e880b?sid=0bf3c689-70ef-4343-ae8f-dd8d492ff49c

🚀 Quick Start

Installation

bash
git clone https://huggingface.co/Agents-MCP-Hackathon/OutageOdyssey
cd OutageOdyssey
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt

Running the MCP server

bash
python mcp_server.py

Running the Gradio App locally

bash
python app.py

Example Incident Input

🚨 Server Outage Incident
Incident ID: INC-2024-001
Title: Production Server Unresponsive
Severity: Critical
Description: Main application server became unresponsive at 14:30 UTC
Impact: 500+ users affected, business operations halted
Symptoms: High CPU usage, memory exhaustion
Actions taken: Attempted restart, no success

See [image] for more examples.

🛠️ Technical Architecture

The project is built on several key components:

  • —Agent Framework: Uses smolagents for multi-step AI reasoning
  • —UI Framework: Custom Gradio interface with comic-themed styling
  • —Image Processing: PIL integration for comic panel generation
  • —File Handling: Support for various incident report formats
  • —Streaming: Real-time agent response streaming

🎨 Supported File Types

Upload your incident documentation in various formats:

  • —.pdf - Incident reports
  • —.docx - Documentation
  • —.txt - Plain text logs
  • —.json - Structured data
  • —.log - System logs
  • —.csv - Data exports
  • —.xlsx - Spreadsheets

🎭 How It Works

  1. 1.Input: Describe your incident or upload documentation
  2. 2.Analysis: AI agent analyzes symptoms and context
  3. 3.Root Cause: Identifies underlying technical issues
  4. 4.Solution: Generates step-by-step resolution recommendations
  5. 5.Story: Creates an entertaining comic narrative
  6. 6.Visuals: Generates comic-style illustrations
  7. 7.Output: Delivers complete educational comic strip

🎪 Example Outputs

The system transforms this:

"Database queries timing out, users reporting slow page loads"

Into this:

🎭 Comic Story: "The Tale of the Grumpy Database Dragon who was hoarding all the memory until the brave DevOps Knight optimized the queries!"

Complete with comic panels showing the "Database Dragon" and resolution steps!

The Tale of the Grumpy Database Dragon

[image]

Some More Example Output Images ....

The Database Resource Exhaustion Case

[image]

The Great Banking BlackOut

[image]

The Epic Server Room Flood

[image]

The Mystery of uneven User flow

[image]

Finally, Issue Resolved

[image]

🌟 Hackathon Integration

This project was specifically designed for Gradio Agents & MCP Hackathon featuring:

🏆 Technologies

  • —HuggingFace - AI model hosting and deployment
  • —Gradio - Interactive web interface framework
  • —MCP - Model Context Protocol integration
  • —Anthropic - Claude AI models for reasoning root cause analysis
  • —OpenAI - GPT models and image generation
  • —smolagents - Multi-step agent framework
  • —Nebius - Cloud infrastructure and compute.
  • —Mistral - Advanced language models and AI reasoning
  • —Hyperbolic - High-performance AI inference platform

🎯 System Features

  • —🤖 AI/ML Innovation - Novel use of agents for IT education
  • —🎨 Creative Tech - Unique combination of technical analysis and storytelling
  • —📚 EdTech - Making technical learning engaging and memorable
  • —🛠️ DevOps Tools - Practical utility for IT professionals

Cloud Deployment

Compatible with major platforms:

  • —Hugging Face Spaces
  • —Google Colab
  • —AWS/Azure/GCP
  • —Nebius Cloud

➕ Enhancements

This project can be extended with:

  • —Additional incident categories
  • —More visual styles
  • —Integration with monitoring tools
  • —Multi-language support
  • —Advanced root cause algorithms
  • —Auto Solution/fix applying tools
  • —Incident to Animated Video generation

📄 License

Open source hackathon project - see individual component licenses.

🎉 Acknowledgments

Special thanks to our hackathon sponsors who made this creative fusion of AI, education, and humor possible:

  • —[HuggingFace](https://huggingface.co/) - AI model hosting and deployment
  • —[Gradio](https://gradio.app/) - Interactive web interface framework
  • —[MCP](https://modelcontextprotocol.io/) - Model Context Protocol integration
  • —[Anthropic](https://anthropic.com/) - Claude AI models for reasoning
  • —[OpenAI](https://openai.com/) - GPT models and image generation
  • —[smolagents](https://github.com/huggingface/smolagents) - Multi-step agent framework
  • —[Nebius](https://nebius.com/) - Cloud infrastructure and compute
  • —[Mistral](https://mistral.ai/) - Advanced language models and AI reasoning
  • —[Hyperbolic](https://hyperbolic.xyz/) - High-performance AI inference platform
  • —[Modal](https://modal.com/) - Serverless cloud computing for AI workloads

🚀 Built for the Gradio Agents MCP Hackathon

Making IT incidents fun to learn from, one comic at a time! 🎭