Image to text (OCR)
Pull the text out of a photo or screenshot.
Runs in your browser. Your files never leave your device.
- Runs fully in your browser
- Confidence score included
- Copy or download the text
How to extract text from an image
- Drop a photo or screenshot here.
- Wait a few seconds while it reads the text.
- Copy or download the result.
What OCR reads well
| Source | Result |
|---|---|
| Clean screenshot of typed text | Excellent, usually 95% or higher confidence |
| Printed document scan | Very good |
| Photo of a sign or page, straight on | Good |
| Photo at an angle, blurry, or low light | Poor, many errors |
| Handwriting | Unreliable, works best on very neat print |
How the confidence score works
The OCR engine returns a confidence percentage for the whole extraction, its own estimate of how sure it is about the text it found. Above about 90, the result is usually accurate. Below 70, read the output against the original image before trusting it, especially for names, numbers and dates that a small misread would change.
Getting a better result
- Crop out anything that is not text before uploading, like a busy background or a photo, since OCR tries to read patterns in everything.
- Straighten a photo taken at an angle first with TabbyKit's Rotate image tool. Tilted text confuses the line detection step.
- Higher resolution helps up to a point. A blurry high resolution photo still reads worse than a sharp lower resolution one.
- Good, even lighting with no glare beats a brighter but unevenly lit photo.
- A flat, straight-on scan of a page almost always outperforms a photo of the same page taken by hand, even from a good camera.
How this OCR engine works
This tool runs Tesseract, an open source OCR engine originally developed at HP and now maintained by Google, entirely inside your browser as WebAssembly. The English text recognition model downloads once, then every photo after that is read on your device with no server involved.
Image to text FAQ
Which languages does this support?
This tool currently reads English text. Results are best on clear, printed or typed text.
Can it read handwriting?
It can sometimes read very neat handwriting, but OCR is built for printed text and struggles with cursive or messy handwriting.
What does the confidence score mean?
It is the OCR engine's own estimate of how sure it is about the text, from 0 to 100. Above about 90 the result is usually accurate; below 70, check it against the original image.
Why does a screenshot read better than a photo?
Screenshots and scans have flat lighting and sharp edges. A photo taken at an angle, with shadows or blur, gives OCR less to work with.
Is my image uploaded to a server?
No. The OCR engine downloads once to your browser, and every photo after that is read on your own device.
Why did it fail to read my image?
This happens rarely, only if the image file itself cannot be decoded. Try a different photo or a direct screenshot instead of a photo of a screen.
Can I copy just part of the extracted text?
Yes. The result appears in an editable text box, so you can select, copy or edit any part of it before downloading.
Does it work on a PDF?
Not directly. Take a screenshot of the PDF page or export it as an image first, then drop that image here.
Can I use this on a phone?
Yes. Everything runs in the browser, including mobile Safari and Chrome, though a very large photo may take a few extra seconds on an older phone.