OCR Online: Image, Handwriting and PDF to Text

Retyping a scanned letter, a photo of a notice board or a handwritten note takes far longer than it should. OCR online (optical character recognition) reads the text in the picture and gives it back to you as text you can edit, copy and search.

Upload a photo, a scan or a PDF and choose how to read it. The AI reader handles handwriting, phone photos taken at an angle and mixed layouts. The fast reader works entirely in your browser and suits clean printed pages. You can then copy the text, listen to it, translate it, or save it as a TXT or JSON file.

Two readers: which one to choose

If the AI reader is busy or unavailable, the page reads the file in your browser instead and tells you so.

Getting the best result

Most OCR errors come from the photo, not the reader. Hold the phone directly above the page, fill the frame with the text, and use daylight or a bright room without your shadow across the paper. For a long document, one clear photo per page reads better than a distant shot of two pages.

The fast reader shows a confidence score. Below about 70%, expect mistakes, and switch to the AI reader. The AI reader gives no score, so check names, numbers and amounts against the original, as you would with any transcription.

Privacy

In AI mode, the file is held in memory only for the request and is not saved. It is still sent to our server and to Google for reading, so for confidential papers with clear printed text, such as bank statements, use the fast reader, which never uploads anything.

How to use this tool

  1. Upload a file: A JPG, PNG or WebP image, or a PDF.
  2. Choose the reader: Best: AI for handwriting and photos; Fast for clean printed text, with its language set.
  3. Extract the text: The text appears in the editor on the right.
  4. Check and correct it: Compare names and numbers with the original; you can edit directly.
  5. Use the text: Copy it, listen to it, translate it, or download it as TXT or JSON.

Features

Worked examples

A handwritten thank-you letter

A phone photo of a one-page handwritten letter, held up by hand and lit by a ceiling lamp.

Fast reader: about 40% confidence, most words garbled.
AI reader: the letter as written, line by line, with an occasional [unclear] word.

Result: For handwriting, the AI reader is the one to use.

A printed notice in Hindi

A clear scan of a printed circular in Hindi.

Result: The fast reader with the language set to Hindi reads it in your browser without uploading it. Translate to English if you need to share it.

A scanned PDF

A three-page scanned agreement with no selectable text.

Result: The fast reader reads every page in turn; the AI reader takes PDFs up to 10 MB. Paste the text into a document to edit it, or download it as a TXT file.

Who it is for

Tips

Frequently asked questions

Can it read handwriting?

Yes, with the AI reader, which is the default. Neat handwriting comes out well; very hurried writing may include a few [unclear] words. The fast reader is not suited to handwriting.

Is my file uploaded?

In AI mode, yes: it is sent to our server and Google's Gemini model to be read, held in memory for the request and not saved. In fast mode nothing is uploaded; the text is read in your browser.

Why is the result full of mistakes?

Usually because the fast reader was used on handwriting or a poor photo. Switch to Best: AI, or retake the photo straight on, closer and in better light.

Which languages are supported?

The fast reader offers English, Hindi, Bengali, Marathi, Tamil, Telugu, Gujarati, French, German and Spanish; choose the language before reading. The AI reader detects the language itself.

Is there a limit?

The AI reader accepts files up to 10 MB and about 20 readings per 15 minutes from one connection. Large phone photos are reduced automatically before upload. The fast reader has no limit beyond your device's memory.

Related tools