Skip to content
OCR2text

Image to text

Extract the text from any image file

Drop in a photo, a scan or an exported page. OCR2text prepares the image, recognises the text in your browser and hands you something you can edit — no upload, no account.

Choose an imageHow the privacy works

How it goes

  1. 1

    Add the image

    Drag it in, pick a file, or paste it. Several images at once is fine.

  2. 2

    Check the page

    Crop, straighten and let Auto enhance clean it up.

  3. 3

    Recognise

    Runs locally in a Web Worker. Cancel any time.

  4. 4

    Correct

    Uncertain words are flagged with the original crop beside them.

  5. 5

    Export

    Copy, or save as TXT, Markdown, DOCX, CSV or JSON.

Which formats work

JPG, PNG, WebP, BMP, GIF and AVIF all decode directly. HEIC photos from an iPhone depend on the browser: Safari on a recent iOS can usually open them, most other browsers cannot. If yours cannot, OCR2text says so plainly and suggests exporting a JPEG rather than silently failing.

Very large images are scaled to a sensible recognition resolution first. Beyond a certain point extra megapixels stop helping the recogniser and start costing memory, so a 48-megapixel phone photo is resized rather than pushed through whole.

Getting a good result from a mediocre photo

Most disappointing OCR comes from the image, not the engine. Fill the frame with the text, avoid glare, and keep the page flat. If the page was photographed at an angle, use Flatten page and drag the four handles onto the corners — the perspective correction usually makes more difference than any other single setting.

For faint or unevenly lit pages, Black & white in Prepare applies an adaptive threshold that handles one side of the page being in shadow. Auto enhance will often pick this for you and tell you it did.

Several pages at once

Add as many images as you like, reorder them in the page rail, then run Recognise all. The text is combined in the order you set. On a phone the queue is deliberately kept to one page at a time so the tab does not run out of memory.

Questions

Is there a file size limit?+
No fixed limit. Very large files are scaled down for recognition, and OCR2text warns you before starting a job that will be heavy on your device.
Does it keep my images?+
Only in the tab's memory while you work. Close the tab and they are gone — nothing is written to a server or a database.
Can I extract only part of an image?+
Yes. Choose Extract selection, drag a box around the part you want, and only that area is recognised.

Extract text from something

No account, no upload, no waiting for a server. Open the workspace and drop a page in.

Choose an image