AI & Agents · Advanced
RAG Over Your Notes
A chatbot that answers from your own PDFs, notes, and Obsidian vault, with sources.
AdvancedweekendPythonAI
Most chatbots answer from generic training data. This one answers from your material. You feed it your PDFs, notes, or an entire Obsidian vault, it embeds and indexes the text, and every answer is generated from the most relevant chunks with citations. This is RAG, the exact pattern behind most production AI products today.
What you build
- Ingest PDFs, markdown, and plain text
- Chunk and embed documents into a vector store
- Semantic retrieval of the most relevant passages
- Grounded answers that cite their sources
- A simple chat UI or CLI
- Swap the LLM between Claude, local, or any OpenAI-compatible model
What it teaches
- Embeddings
- Vector search
- Retrieval-augmented generation
- Chunking strategies
- Grounding and citations
- LLM APIs
How it works
- 1
Your docs
- PDF, notes, Obsidian
- 2
Chunk + embed
↓ vectors
- 3
Vector DB
- ChromaDB
- 4
Retrieve
↓ top-k chunks
- 5
LLM answer
- Claude, with citations
Sign in to open the build guide
Free account. Get the step-by-step build and every resource link.
Take it further
- Add hybrid search that combines keyword and vector retrieval.
- Re-rank retrieved chunks for higher precision.
- Add conversation memory so follow-up questions keep context.
