Chat with PDF

Summarize and question any document, privately

Runs entirely in your browser. Your document is never uploaded, stored or used for training.

Frequently asked questions

Is my document uploaded to a server?

No. The PDF is opened, read and searched entirely inside your browser by our own engine. No copy of your file is sent anywhere, nothing is stored, and nothing is used to train a model.

How is this free?

Because the work happens on your device instead of ours. There is no server doing the reading, so there is no per-document cost to pass on. The small search model is downloaded once from a public model host and then cached by your browser.

Do I need a fast computer or a special browser?

No. Summarizing and searching use a compact model that runs on any modern browser without a graphics card, and it only downloads once.

Does it work on scanned PDFs?

Only if the scan already has a text layer. A photo-only PDF has no text to read, so run it through OCR first to make it searchable, then come back here.

Are the answers made up?

No. Every summary line and every search result is quoted directly from your document and labelled with the page it came from, so you can always check it against the source.