Chat with PDF
Upload up to 10 PDFs and ask anything — grounded, streaming answers with automatic OCR.
Chat
Upload PDFs to start
Ready to Chat
Upload PDFs to start
Up to 10 PDFs
Chat across a whole set of documents in one conversation.
Streaming Answers
Responses stream as the AI thinks — no waiting for the full reply.
Auto-OCR
Scanned PDFs are OCR'd on-device before they hit the chat.
Local Extraction
PDFs never leave your browser — only extracted text is sent to the AI.
Frequently Asked Questions
How does Chat with PDF work?
Text is extracted from each PDF locally in your browser using pdf.js. If a PDF has no selectable text (it's a scan), Tesseract OCR runs in-browser to recognise the text. The extracted text is then sent to the AI which answers your questions grounded strictly in the documents.
Can I chat with multiple PDFs at once?
Yes — upload up to 10 PDFs. The AI sees them as separate labelled documents and can cross-reference between them.
Are my PDFs uploaded anywhere?
No. The PDF bytes stay on your device. Only the plain text extracted from them is sent to the AI provider for question answering.
What about scanned PDFs?
When no text layer is detected, the app runs Tesseract OCR at 200 DPI in your browser, then feeds the recognised text into the chat.
Are my chats saved?
No — conversation history stays in this browser tab and is cleared when you refresh or navigate away.
What are the size limits?
There's no hard PDF size cap, but the combined extracted text is truncated to 200,000 characters before sending to the AI to keep answers focused.