Document RAG
Chat with your notes using a RAG pipeline
Network disclosure: Runs locally by default; calls the Hugging Face API only if you add an HF token
About Document RAG
How Document RAG Works
Text is chunked in the browser, ranked by relevance, then answered extractively or with Flan-T5 when a HF token is present.
Document RAG lets you paste notes or specs, chunk them, retrieve relevant passages, and answer questions — a hands-on RAG demo that still respects optional privacy controls.
Frequently asked questions
Can I upload a PDF?+
Paste extracted text or import a .txt/.md file. Binary PDF parsing is not included in this version.
Is this the same as ChatGPT?+
No. It is a retrieval pipeline over your pasted document, optionally using Hugging Face models you authorize.
Limitations
Answer quality depends on chunk size and document clarity. Always verify critical answers against the source text.