PDF Source Prep

Stop dumping whole PDFs into ChatGPT.

Pick just the pages you need from your PDFs, bundle them with a citation-ready prompt, and hand a clean source packet to Claude or ChatGPT — figures and layout kept intact.

100% in your browserNothing uploadedFree
or drag them into the workspace below

Drop PDFs here, or click to add

Add as many as you like · processed only in your browser, never uploaded

0 files · 0 pages → 0 selected

How it works

1

Load your PDFs

Add as many as you like. Everything is processed in your browser — files never leave your device.

2

Keep only what matters

Search across text PDFs or pick pages by hand, zooming in to check each one before you commit.

3

Export a packet

Download one .zip: trimmed PDFs plus a citation-ready prompt to paste straight into Claude or ChatGPT.

Why not just upload the whole PDF?

Big or many PDFs overflow the context window.

A few full books won't fit. Trimming to the pages that matter keeps everything inside the model's limit.

Noise dilutes the model's attention.

Hand it 200 pages and it reads the 190 you didn't need. A focused packet keeps the answer on target.

Citations stay straight.

Ask a model about a raw upload and it can only point to "page 34 of the file" — or invent one. The included prompt tells it to cite by source and read the printed page number, and never to guess.

Guides

What to do when ChatGPT says your PDF is too longWhy it happens, and how trimming to the right pages fixes it. Why ChatGPT invents page numbers from your PDFsThe one prompt that keeps AI citations honest and checkable. How to feed multiple PDFs to Claude or ChatGPTGetting several sources in without hitting the file or context limit.