English only for now
The bundled OCR model recognizes English. Other languages may be partially readable but are not a supported result.
Image tools
Make a picture of words into words.
Extract text from screenshots, receipts, signs, and scanned pages. The OCR engine runs in your browser, and the text stays on your device.
No upload. No account. Files stay on your device. The processing happens in this browser tab—an unusual arrangement in a category full of upload forms.
Add images
0 of 1 · 25.0 MB eachDrop images here, or choose files
One JPG, PNG, or WebP file, up to 25.0 MB.
Added locally. Nothing is uploaded.
Recognise the text
English recognition works best on straight, high-contrast text. Screenshots are reliable; photos with glare, skew, handwriting, or tiny print need proofreading. The image itself is never uploaded.
Add a screenshot, scan, or photo of text.
Your image is decoded and re-encoded by this browser. There is no upload step, no server-side queue, and no stored copy. How local processing works.
The short version
How it works.
No installer, no sign-up, and no second app to open. The work happens in the browser you already have, using the file already on your device.
-
Add an image
Choose a clear, reasonably straight image. Higher resolution and good contrast give the recognizer more to work with.
-
Run recognition
Start OCR and wait for the language model to load on first use. The text is produced by a local recognition engine.
-
Check and copy
Review the output for misread characters, then copy it or save a .txt file. Keep the image if the text is legally or financially important.
Fine print, without the runaround
Good to know.
Every tool has a boundary. These are the ones worth knowing before you start, so the result is not a surprise.
The first run downloads the engine
OCR needs a small language model and a worker script. They are fetched once and may be cached by the browser; the image itself is never uploaded.
How accurate is the OCR?
Clean screenshots recognize very well. Photos with glare, skew, handwriting, or tiny print are less reliable. Review names, amounts, and reference numbers manually.
Does it work offline?
After the OCR assets have been loaded once, the browser may cache them and continue to work offline. On the first use a connection is needed to fetch the engine.
Can I extract text from a PDF instead?
Yes. Use PDF text & OCR, which handles selectable text first and can also recognize scanned PDF pages.
Next door
Related tools.
Most file jobs come in pairs. These three usually follow close behind.