README.md

metadata

title: Gradio Docs Chat
emoji: 🤖
colorFrom: blue
colorTo: purple
sdk: gradio
sdk_version: 5.44.1
app_file: app.py
pinned: false

🤖 Gradio Docs Chat

A powerful chat interface for Gradio documentation, powered by OpenAI GPT-OSS-20B with MCP (Model Context Protocol) integration.

✨ Features

• 🧠 GPT-OSS-20B: Advanced reasoning with harmony format
• 🔧 MCP Integration: Real-time access to Gradio documentation
• 📊 Live Tool Logs: See what tools are being used
• 📎 Citations: Track sources and references
• ⚙️ Answer Styles: Choose between concise or detailed responses
• 🌊 Streaming: Real-time response streaming

🚀 How to Use

1. Ask Questions: Type your Gradio-related questions
2. Choose Style: Select "Concise" or "Detailed" response style
3. Watch Tools: See live tool activity in the side panel
4. Check Sources: View citations and references

🔧 Configuration

Space Secrets (Optional)

Set these in your Space settings for customization:

• HUGGING_FACE_HUB_TOKEN or HF_TOKEN: For model access
• CHAT_MODEL: Model is hardcoded to openai/gpt-oss-20b
• CHAT_PROVIDER: Model provider (default: auto)
• GRADIO_DOCS_MCP_SSE: MCP server URL

🏗️ Architecture

• Frontend: Gradio Blocks with streaming chat interface
• Backend: GPT-OSS-20B with transformers pipeline
• MCP: Gradio Docs MCP server for documentation access
• Deployment: Hugging Face Spaces

📦 Dependencies

• gradio>=5.0.0: Web interface
• transformers>=4.40.0: Model loading and inference
• huggingface_hub>=0.34.0: MCP client
• torch>=2.0.0: PyTorch backend
• accelerate>=0.20.0: Model acceleration

🎯 Example Questions

• "How do I create a simple Gradio interface?"
• "What are the different input components available?"
• "How can I add custom CSS to my Gradio app?"
• "Explain the difference between Blocks and Interface"
• "How do I deploy my Gradio app to Hugging Face Spaces?"

🔗 Links

• Gradio Documentation
• GPT-OSS-20B Model
• MCP Protocol

📄 License

This project is open source and available under the Apache 2.0 license.