How karthick-005/multi-url-rag-chatbot works
This is a single-file, in-memory multi-URL RAG chatbot. Users submit URLs through Gradio, which dispatches ingestion to fetch and clean pages, chunk text, and build a FAISS index with embeddings. Questions go through the chat handler to retrieval and context-based answer generation via the Hugging Face Router. The graph reflects calls shown in app.py and capabilities documented in README; distinct implementation stages share that source path.
Main entry points: app.py URL processing
Components 14
Chat interface
- Gradio UI
app.py - Chat handler
app.py
URL ingestion
- URL processing
app.py - Document loader
app.py - HTML cleaning
app.py - Text chunking
app.py - Embeddings
app.py
Retrieval and answers
- FAISS index
app.py - Question answering
app.py - Context assembly
app.py - LLM request
app.py
Other components
- User
- Web pages
- Hugging Face Router
Connections 14
- User → Gradio UI: uses
README.md - Gradio UI → URL processing: dispatches
app.py - URL processing → Document loader: loads and chunks
app.py - Document loader → Web pages: fetches
app.py - Document loader → HTML cleaning: cleans
app.py - Document loader → Text chunking: splits text
app.py - URL processing → Embeddings: uses
app.py - URL processing → FAISS index: builds index
app.py - Gradio UI → Chat handler: dispatches
app.py - Chat handler → Question answering: calls
app.py - Question answering → FAISS index: searches
app.py - Question answering → Context assembly: assembles
app.py - Question answering → LLM request: requests answer
app.py - LLM request → Hugging Face Router: sends prompt
app.py