RAG Chat

Retrieval-augmented chat over text you paste in — embedded, chunked, and retrieved per question.

Paste text to index

Document text

Chunks retrieved last query

Uses an in-memory vector store by default (see lib/vector-store.ts). Swap in Pinecone/Supabase for persistence across restarts.

Answers are grounded in your indexed text only

Index a document, then ask about it

Paste text on the left, queue it, then ask a question here.