Free tools › Images
OCR with coordinates
Not just the text: where each word sits on the page and how sure the engine was, grouped into lines — for forms, receipts and anything you parse afterwards.
More free conversions, where you upload. The Fylane extension does the same work inside any upload form, on its own free share every day.Add to Chrome
How to ocr with coordinates
- Drop the image.
- Pick the language.
- Press Read layout and download the JSON.
What happens to your file
The image is straightened and cleaned, then read with the geometry kept: every word arrives with its box in the image’s own pixels, its confidence from 0 to 100, and the line it belongs to. The plain text comes along as a second file for anyone who only wanted the words.
This is the answer for a receipt whose total you want to find by position, a form whose fields sit in known places, or a scanned table you rebuild yourself. Image to text gives you the words; this page tells you where they were.
A scanned PDF is not read directly — render its pages to images first with PDF to JPG, then read those.
Questions
What is in the JSON?
A list of lines, each with its words; every word has text, a box (x, y, width, height in pixels) and a confidence. The file also carries the word count and the mean confidence, so you can tell a clean scan from a blurry photo before parsing.
Can I get the text alone?
It comes with every run as a second file — and Image to text is the page for that alone.
Are the images kept?
Deleted after 24 hours; the JSON after 7 days. Nobody reads either.
Do it once, not on every form
A form that wants a square JPEG under 240 KB says so on the page. The Fylane extension reads that, fits your photo to it, and puts it in the field — without you working out what 240 KB looks like.
And it keeps the result. The next application that asks for a résumé, a cover letter and a photo fills all three from one tap, each one refitted to what that site accepts.
Your files stay in the browser, encrypted, on your machine. Nothing is uploaded to keep them and nothing is sent anywhere to fill a form.
Related tools
Do this in code
curl -X POST https://api.fylane.dev/v1/jobs \
-H "Authorization: Bearer $FYLANE_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"file_id": "<file id>",
"operation": "ocr.layout",
"options": {
"language": "eng"
}
}'The same operation, ocr.layout — OCR that keeps the geometry: every word with its box and confidence, grouped into lines. Upload first with the quickstart; every option is on the operations page. 500 free credits a month, then $5 per thousand.