server/README.md

Podcast Discussion Server

This is a FastAPI-based server that provides podcast discussion and analysis capabilities.

Environment Variables

The following environment variables need to be set in the Hugging Face Space:

• OPENAI_API_KEY: Your OpenAI API key
• ALLOWED_ORIGINS: Comma-separated list of allowed origins (optional, defaults to Vercel domains and localhost)

API Endpoints

• /chat: Main chat endpoint for podcast discussions
• /podcast-chat/{podcast_id}: Chat endpoint for specific podcast discussions
• /audio-list: List available audio files
• /audio/{filename}: Get specific audio file
• /podcast/{podcast_id}/context: Get podcast context

Stack

• FastAPI
• OpenAI
• Langchain
• Qdrant
• GTTS

Deployment

Backend (Hugging Face Spaces)

This server is deployed on Hugging Face Spaces using their Docker deployment feature.

Frontend (Vercel)

When deploying the frontend to Vercel:

1. Set the API base URL in your frontend environment:

   VITE_API_BASE_URL=https://your-username-your-space-name.hf.space
   

2. The server is already configured to accept requests from:

   • All Vercel domains (*.vercel.app)
   • Local development servers (localhost:3000, localhost:5173)

3. If you're using a custom domain, add it to the ALLOWED_ORIGINS environment variable in your Hugging Face Space:

   ALLOWED_ORIGINS=https://your-custom-domain.com,https://www.your-custom-domain.com
   

Security Features

• CORS protection with specific origin allowlist
• Security headers (HSTS, XSS Protection, etc.)
• Rate limiting
• SSL/TLS encryption (provided by Hugging Face Spaces)