ChatGPT Can't Read Your PDF? Get the Text Out
When ChatGPT says “No text could be extracted from this file”, your PDF is usually a scan. Drop it here and every page becomes text you can paste into ChatGPT.
Drop your PDF here
Scanned or normal. Nothing is uploaded.
Opening…
Scanned pages were read by text recognition, so skim them for mistakes before you rely on numbers or names.
- Free
- No sign-up
- No ads
- Runs in your browser, nothing is uploaded
How to fix “No text could be extracted from this file”
Drop the PDF above
Pages that already contain text are read instantly. Scanned pages are recognized one by one; you'll see which page it's on.
Copy the text
Press “Copy text”. Long PDFs are split with “--- Page N ---” markers so ChatGPT can tell you which page a quote came from.
Paste it into ChatGPT
Paste the text into the chat, or save it with “Download .txt” and upload that file instead. Then ask your question as usual.
Why ChatGPT can't extract text from some PDFs
A PDF can hold two very different things. Most PDFs exported from Word or Google Docs contain real text, and ChatGPT reads them fine. A scanned PDF, or one made from phone photos, contains only pictures of pages. There are no letters inside the file for a text reader to find, so ChatGPT reports that no text could be extracted.
Less often, the PDF has text but it's stored in a way readers can't decode (some older or print-ready PDFs do this), or the file is password-protected. For the first case, tick “Recognize every page” above. For the second, save a copy without the password first.
Other ways to get ChatGPT to read a scanned PDF
| Way | Good for | Downside |
|---|---|---|
| Extract the text here, then paste it | Any number of pages; you can check the text first | Layout and images are dropped |
| Screenshot each page and upload the images | One or two pages, or pages with charts | Slow for long documents; each image counts toward upload limits |
| Run OCR in a desktop PDF app, then upload the PDF again | Keeping the PDF for later | Usually needs a paid app or an upload to someone's server |
What it won't do
- It doesn't keep the layout. Tables come out as lines of text, and two-column pages may mix the columns.
- It can misread. Recognition of scanned pages is good on clean printed text and worse on faint, skewed or handwritten pages. Check names and numbers.
- It doesn't talk to ChatGPT for you. You copy the text and paste it in yourself; nothing is sent to OpenAI from this page.
FAQ
Why does ChatGPT say “No text could be extracted from this file”?
Most of the time the PDF is a scan: each page is a picture of text, not text. ChatGPT's file reader looks for the text inside the PDF, finds none, and gives up.
Is my PDF uploaded anywhere?
No. The PDF is read and recognized in your browser. The reader and the language file are downloaded to your browser, but your file never leaves the page.
How long does it take?
Pages that already contain text take a moment. Scanned pages take a few seconds each, and the first scan downloads a language file of about 10 MB.
Does it work with handwriting?
Not well. It's built for printed text. Neat handwriting may partly come through; messy handwriting won't.
Which languages are supported?
English, Spanish, French, German, Portuguese, Italian, Chinese (Simplified) and Japanese. Pick the language before you drop the file.