AI Docs Assistant
Internal Demo · Code Available · No Live DemoA retrieval-based chat assistant that answers from a company's own documentation and shows the source of every answer. Python and FastAPI back end, Next.js front end, works with Claude, OpenAI or offline.
Problem
Support and documentation bots often sound confident while making things up, and nobody can see where an answer came from. Businesses need an assistant that stays inside their own content and admits when the answer isn't there.
Solution
Built a retrieval-based assistant. The Python back end splits documents by heading, finds the passages that match a question with BM25 search, and streams the answer back as it is written, while the Next.js front end shows clickable citations beside every answer. A question with nothing relevant in the docs never reaches the model and gets an honest "could not find it" reply instead of a guess.
Technology
Key Features
- Streams answers live with clickable source citations
- Relevance check keeps off-topic questions away from the model
- Works with Claude, OpenAI, or an offline mode that needs no API key
- Input validation, rate limiting and server-side API keys
- 58 automated tests, including a retrieval accuracy check