Apppublic

NaimalcreativityAI/physical-ai-chatbot-api

sourceHugging Facemitupdated 10mo agoView on Hugging Face

0likes

App README

Physical AI Chatbot API 🤖

RAG-based chatbot for the Physical AI & Humanoid Robotics textbook.

Features

—✅ Answer questions from textbook content exclusively
—✅ Text selection queries - Ask about highlighted text
—✅ Navigation guidance - Get links to relevant lessons
—✅ Study companion - Recommendations for learning path
—✅ Streaming responses via Server-Sent Events (SSE)
—✅ Source citations - Every answer includes chapter/lesson references

Technology Stack

—Framework: FastAPI
—LLM: Google Gemini 1.5 Flash (free tier)
—Embeddings: Qdrant FastEmbed (BAAI/bge-small-en-v1.5)
—Vector DB: Qdrant Cloud (free tier)
—Streaming: SSE (Server-Sent Events)

API Endpoints

Health Check

bash

GET /api/health

Chat

bash

POST /api/chat
Content-Type: application/json

{
  "message": "What is embodied intelligence?",
  "selected_text": null,
  "current_page": null,
  "conversation_history": []
}

Interactive API Documentation

—Swagger UI: /docs
—ReDoc: /redoc

Environment Variables

Required secrets (configure in Hugging Face Space settings):

—GEMINI_API_KEY - Get from https://aistudio.google.com/app/apikey
—QDRANT_URL - Qdrant Cloud cluster URL
—QDRANT_API_KEY - Qdrant Cloud API key
—CORS_ORIGINS - Allowed CORS origins (e.g., https://yoursite.com)

Local Development

bash

# Install dependencies
pip install -r requirements.txt

# Set environment variables
export GEMINI_API_KEY=your_key
export QDRANT_URL=your_qdrant_url
export QDRANT_API_KEY=your_qdrant_key

# Run server
python app.py

Frontend Integration

This API is designed to work with the Physical AI textbook frontend:

—Frontend: https://naimalarain13.github.io/physical-ai-and-humaniod-robotics/
—Repository: https://github.com/NaimalArain13/physical-ai-and-humaniod-robotics

Cost

💰 Total Cost: $0/month

—Gemini API: Free tier (15 requests/minute)
—Qdrant FastEmbed: Local, free
—Qdrant Cloud: Free tier (1GB storage)

License

MIT License

Support

For issues and questions, please visit the GitHub repository.