AZTools.us
English

Image to Text (OCR) – Photos & Scanned PDFs

Image beside the recognised text 12 languages · scanned PDFs too

Public-domain text, typeset for this site, downloaded to your browser only when you click.

Image to Text (OCR) settings

Load a file to unlock this tool’s controls:

Image beside the recognised text

Runs directly on the web inside your browser. Your original file is never modified or uploaded.

Retyping text from a photo, a screenshot or a scanned document is slow. This tool reads the text for you with Tesseract, an open-source OCR engine, running in your browser. It works on images and PDFs, in twelve languages, and lets you copy the result or save it as a .txt file. Files are not uploaded.

Updated

How to convert an image to text

  1. Drop .jpg, .jpeg, .png, .webp, .bmp, .gif or .pdf files onto the workspace. Use “+ Add files” to add more to the list.
  2. Check the “Language”: it starts with the language of this page. If the text mixes two languages, pick the second one under “Second language”, for example Vietnamese with English.
  3. Wait for each row to finish. The first run downloads the OCR engine (about 4 MB) and the language data (1–3 MB per language) from aztools.us; later runs start straight away.
  4. Select a file to see it beside its text. Press “Copy text” to put the text on the clipboard.
  5. For PDFs, “Use the PDF’s own text” is on by default: pages that already contain text are copied exactly and instantly, and only scanned pages go through OCR. Turn it off to read every page as an image.
  6. Export to save the text as a .txt file. Several files are saved together in one .txt, each under its file name.

Languages

English, Vietnamese, Chinese (Simplified), Chinese (Traditional), Japanese, Korean, Spanish, Portuguese, German, French, Italian and Russian. Two languages can be read together; each extra language makes recognition a little slower.

Getting good results

  • Use a sharp, evenly lit photo taken straight on, with the text upright. Small images are enlarged automatically, but blur cannot be undone.
  • Printed text works best. Handwriting, decorative fonts and text on busy backgrounds are often misread or missed.
  • Each row shows how many characters were found and Tesseract’s average confidence; below 60% the row is marked so you know to check the text.
  • Tables and columns come out as plain lines of text; the layout is not kept.

Image to Text (OCR): common questions

Can it read a scanned PDF?
Yes. Each page without a text layer is rendered at about 300 dpi and read by OCR; the result shows the text page by page. Long PDFs take a while, and the progress line shows which page is being read.
Why is some text wrong?
OCR guesses each character from its shape, so blur, low contrast, unusual fonts or the wrong language setting cause mistakes. Check the language, try a sharper photo, and correct the remaining errors after copying.
Are my files uploaded?
No. Recognition runs in your browser tab. The OCR engine and language data are downloaded from aztools.us, once, and your images and PDFs are never sent anywhere.