PDF to Text Free: Scans Work, but a Run Stops at 20 Pages

Open the tool at https://avaloncompany.ai/tools/pdf-to-text/. Drop your PDF onto the page. If your PDF is a scan, tick the language of the scan first. English is ticked already.

Press Extract text. The text of every page appears in a box. You can copy it, fix anything you need, or save it as a .txt file.

Each line of text becomes a line in the result. In a multi-page PDF, each page starts with its page number.

The tool uses the PDF's own text where it exists. It uses OCR (optical character recognition) only for pages that are pictures, such as scans and phone photos saved as PDF. If the copied text comes out as strange symbols or letters in the wrong order, choose OCR on every page. Some PDFs store their text in a broken way, and reading the page as a picture can give a better result.

The PDF is read inside your browser, on your own device. Nothing is sent to our servers or anyone else.

On 2026-11-01 we ran 16 test cases in Chromium (desktop) on the free tier. All 16 passed. Here, a pass means the tool did the right thing for that file. For a good file, that means it produced text. For a bad file, it means it refused with a clear message.

The cases fall into three groups. One normal case used a typical file. Twelve edge cases covered unusual file names, damaged files, a locked file, mixed page sizes, many scripts and a scan. Three limit cases pushed the free size and file limits, and one more pushed the page limit.

The results table on this page lists every case. The next section covers the ones that were hard.

Files that could be read worked. A very long file name, a name with extra spaces and an emoji, and a name with Korean and Japanese characters all ended with "Done: text from 1 file(s)". So did a PDF with pages of different sizes. So did a PDF with Korean, Japanese, Chinese, Arabic, Hindi, Thai and Russian text, and a scanned, image-only PDF. We checked that each run finished and returned text. We did not grade how accurate the OCR was.

Files that could not be read were stopped with a clear message. A 1-byte file, a 0-byte file, a cut-off (truncated) PDF and a JPEG renamed to look like a PDF all gave: "This file is not a valid PDF or it is damaged." Renaming a picture to .pdf does not turn it into a PDF.

A password-protected PDF gave: "This PDF is locked. Unlock it in your PDF reader, save an unlocked copy, then add that copy." So unlock the file first, then add the copy.

In our tests the free version held at most 2 PDFs in a batch. Two files were accepted with the message "2 file(s) ready". A third file gave: "This batch is full (2 PDF at most). Extract the text first, press Clear, then add the next."

A batch also holds 25 MB in total. When we went 1 MB over, the tool said: "pdf_mb_26.pdf was not added: one batch can hold 25 MB in total. Extract these first, press Clear, then add the rest."

A run reads up to 20 pages. With 21 pages, the tool said: "Stopped after 20 PDF pages in one run. The text so far is below; split the PDF and run the rest separately." You keep the text from the first 20 pages, so split the file and run the rest.

Avalon Pro reads up to more PDFs and more pages in one run, and adds a ZIP download, saved presets and your own file names. If your PDF is locked, damaged or not really a PDF, fix the file first. This tool will not repair it.