Is this really free and unlimited?

Yes. The recognition engine runs in your browser, so there is no server cost to pass on and no page limit.

Why is the first run slow?

The engine (about 4 MB) and the language file (2–15 MB) are downloaded into your browser the first time. After that they are cached.

Printed text in a clear photo reads at close to 100%. Handwriting, tiny text and blurry pictures will contain mistakes — check the result before relying on it.

Yes. Choose Hindi, or English + Hindi for mixed documents. Marathi, Bengali, Tamil, Telugu, Gujarati, Kannada, Malayalam, Punjabi and Urdu are also available.

Which languages are supported?

Twenty in total: English, Hindi, Marathi, Bengali, Tamil, Telugu, Gujarati, Kannada, Malayalam, Punjabi, Urdu, Arabic, Spanish, French, German, Portuguese, Russian, simplified Chinese and Japanese — plus an English + Hindi mode for mixed documents. Pick the right one before you start; the wrong language is the most common reason for gibberish output.

What does the result look like?

Plain text in reading order, which you can copy or download. Layout, fonts, columns and tables are not preserved — it reads the words, not the design.

Why is a whole paragraph missing?

The engine skips regions it cannot make sense of: text over a photo, watermarks behind the words, heavy skew or a deep crease. Crop tightly to the text block and straighten the photo, then try that part again.

Does the scanned PDF get uploaded?

No. Pages are rendered and read entirely on your device.

Convert your files in seconds

Free image to text converter (OCR)

Extract text from a photo, screenshot or scanned PDF in your browser. English, Hindi and 18 more languages. Free, nothing uploaded.

Drop files here

or browse your device — one file or many; images, documents, spreadsheets, audio, video, archives, 3D models

Read the words out of a picture or a scan and copy them as text. The recognition runs on your device, so a private document stays private. Nothing is uploaded: the work happens inside your browser, on your own device.

How to use it

Drop images or a scanned PDF Straight, sharp, well-lit pictures read best.
Choose the language English is the default; Hindi and other Indian languages are available.
Read and copy The text appears in a box you can copy from or download.

Questions

Is this really free and unlimited?: Yes. The recognition engine runs in your browser, so there is no server cost to pass on and no page limit.
Why is the first run slow?: The engine (about 4 MB) and the language file (2–15 MB) are downloaded into your browser the first time. After that they are cached.
How accurate is it?: Printed text in a clear photo reads at close to 100%. Handwriting, tiny text and blurry pictures will contain mistakes — check the result before relying on it.
Can it read Hindi?: Yes. Choose Hindi, or English + Hindi for mixed documents. Marathi, Bengali, Tamil, Telugu, Gujarati, Kannada, Malayalam, Punjabi and Urdu are also available.
Which languages are supported?: Twenty in total: English, Hindi, Marathi, Bengali, Tamil, Telugu, Gujarati, Kannada, Malayalam, Punjabi, Urdu, Arabic, Spanish, French, German, Portuguese, Russian, simplified Chinese and Japanese — plus an English + Hindi mode for mixed documents. Pick the right one before you start; the wrong language is the most common reason for gibberish output.
What does the result look like?: Plain text in reading order, which you can copy or download. Layout, fonts, columns and tables are not preserved — it reads the words, not the design.
Why is a whole paragraph missing?: The engine skips regions it cannot make sense of: text over a photo, watermarks behind the words, heavy skew or a deep crease. Crop tightly to the text block and straighten the photo, then try that part again.
Does the scanned PDF get uploaded?: No. Pages are rendered and read entirely on your device.

All tools · All 399 conversions

My files

Everything you have converted, kept in this browser. Download any of them again, any time.

Formats

Every conversion this browser can do. Pairs that would not genuinely work are simply not offered.

0conversion pairs

0source formats

0target formats

Guides

Practical, plain-language answers about file formats and the conversions people ask about most.