Image to Text
Pull the text out of a screenshot, photo or scan.
This image is sent to a server to be read. If that is wrong for a particular document, don't use this one.
Drop a screenshot or photo here
Printed text, handwriting, signs, receipts. Up to 8MB — larger images are scaled down on your device before anything is sent.
Reading the text…
That didn't work.
This browser can't display that format, so the file was sent as it is and read on the server. The text is below.
This tool sends your image to a server to be read. It is held in memory for the second or two that takes, never written to disk, and never kept afterwards.
Keyboard shortcuts
- / or ⌘K
- Search every tool
- ⌘↵
- Run this tool, from anywhere on the page
- ⌘S
- Download the result
- ?
- Open and close this list
About the Pathine Image to Text
Pulls the text out of a screenshot, a photo or a scan and gives it back as text you can copy, edit and search. Printed pages, handwriting, receipts, signs, slides — anything with words in it.
This is one of the few tools on Pathine that does not run on your device. Reading text out of a photograph well needs a model far too large to download, so the image is sent to a server, read, and the text comes back. It is held in memory for the second or two that takes, never written to disk and never kept. The rest of the site is built the other way round, and the tool says which it is before you hand it anything.
Your image is scaled down on your device before it is sent — enough resolution to read small print, far less than a phone camera produces. That makes it quicker and means less of your document crosses the network than you gave it.
When you'd reach for it
Text trapped in a screenshot
An error message, a slide, a page someone photographed rather than copied. Getting the words out beats retyping them and beats reading them aloud to someone.
A printed document you need to edit
A contract, a letter or a form that only exists on paper. Photograph it and you have editable text in a couple of seconds.
Receipts and invoices
Pulling the totals and dates off a receipt is far faster than typing them into a spreadsheet, and less error-prone at the end of a long day.
Handwritten notes
Notebook pages, whiteboards, a list on the back of an envelope. Handwriting is where traditional OCR falls apart and where a modern model is genuinely good.
How to get text out of an image
-
Add your image
Drop it, paste it, or choose a file. JPG, PNG, WebP, HEIC and AVIF all work, up to 8MB.
-
Wait a second or two
The image is scaled down on your device, sent, read and returned. Typically about a second.
-
Read and correct
The text appears in an editable box. It is very accurate but not infallible — worth a glance before you rely on it, especially for numbers.
-
Copy or download
Take it to the clipboard, or save it as a text file. Nothing is kept here once the response is sent.
Common questions
Does this one really send my image to a server?
Yes, and only a handful of tools here do. Reading text out of a photograph needs a model far too large to run in a browser tab, and the same is true of the writing and description tools. Everything else on Pathine — compressing, converting, merging, stripping metadata, even removing a background — genuinely runs on your device. The pages that are different say so at the top.
Is my image stored?
No. It exists in memory for the second or two the request takes and is never written to disk or to a database. We keep a count of how many extractions have happened, tied to a one-way hash of your address purely to enforce a daily limit — not the image and not the text.
How accurate is it?
Very, on printed text — near-perfect on clean pages, and it holds up on blurred, noisy, skewed and handwritten input where older OCR tools fail badly. It is not infallible. Check numbers before relying on them: a misread digit in an amount is the error that actually costs something.
Is there a limit on how much I can use it?
Yes — a daily allowance per visitor, and the tool tells you what is left. Each extraction costs us money, and a free tool with no account needs a ceiling or it becomes somebody's free API. It is generous for a person and useless for a script.
What languages does it handle?
A great many, including scripts that are not Latin. The text comes back in the language it was written in — this transcribes rather than translates.
Can it read a PDF?
This one takes images. PDF to Text is the tool for a document: it reads the text a PDF already contains on your device, without sending anything, and only falls back to this kind of recognition for pages that turn out to be scans — and only for those pages.