Turn any image into text you can edit
Drop in a screenshot, a scan, a receipt or a photo of a whiteboard. We read the words with OCR, show you how confident we are about every word, and hand back editable text — as TXT, DOCX or a searchable PDF. No retyping, no watermark, no guessing about accuracy.
Drop images here, or click to choose
JPG, PNG, WebP, GIF, BMP, TIFF — screenshots, scans, receipts, whiteboards, handwritten notes
How it works
1. Upload
Add one image or a whole batch. JPG, PNG, WebP, GIF, BMP and TIFF all work — screenshots, scans, receipts, whiteboards and handwritten notes.
2. Pick a language and format
Choose the text language and how you want the result: plain TXT on every plan, DOCX, searchable PDF or JSON with word positions on paid plans.
3. Check the confidence report
Every page reports mean confidence, and words we are unsure about are flagged for proofreading. If Tesseract comes back nearly empty, paid plans re-read the page with a vision model.
4. Download
Results come down as one ZIP with the text files and an OCR report. Registered users get 72-hour history and an email when big batches finish.
What you can convert
| From | Typical source | Note |
|---|---|---|
| JPG to text | Photos and phone screenshots | The most common case — a photo of a page, a whiteboard or a sign |
| PNG to text | Screenshots and scans | Clean edges read best; PNG screenshots are usually near-perfect |
| Screenshot to text | Anything on your screen | Error messages, chats, tables and forms you cannot copy |
| Receipt to text | Expenses and bookkeeping | Totals and line items as editable text, not a photo |
| Handwriting to text | Notes and whiteboards | Printed handwriting reads well; cursive is harder — we say so in the report |
| Scanned PDF pages to text | One image per page | Export each page as an image first, then convert the batch |
Built for people who hate retyping
Batch the whole folder
Starter converts 50 images at once, Pro 200. A semester of lecture slides or a drawer of receipts becomes one ZIP of text files.
Proofread with confidence
We never pretend a guess is certain. Low-confidence words are listed in the report and in the DOCX, so you check ten words instead of re-reading the whole page.
Searchable PDFs
Paid plans return a PDF that keeps your original image and adds a hidden text layer — search it, copy from it, archive it.
Private by default
Your images are processed on our server, never shared, and deleted automatically after 72 hours.
Frequently asked questions
How accurate is the OCR?
On clean printed text — screenshots, typed documents, sharp photos — Tesseract typically reads 95% or more of the characters correctly, and we show you the mean confidence for every page so you can judge for yourself. Blurry photos, skewed scans and handwriting score lower, and words we are unsure about are flagged for proofreading. Paid plans add a vision-model re-read for pages where Tesseract is unsure, which recovers many pages that would otherwise come back nearly empty.
What happens to my images? Do you keep them?
Uploads and results are deleted automatically after 72 hours. Files are processed on our own server and are never shared or used to train models. See the privacy policy for the details.
Which image formats work?
JPG, PNG, WebP, GIF, BMP and TIFF. JPG and PNG cover almost everything; WebP works too. If a file is corrupt or in another format, the job tells you so instead of charging you for it.
What is the difference between the free and paid plans?
Free gives you 10 pages a month, one image at a time, with TXT export. Starter ($9/month) raises that to 200 pages, 50 images per batch, adds the high-accuracy re-read and DOCX, searchable PDF and JSON export. Pro ($19/month) goes to 1,000 pages, 200 images per batch, larger files and API access. Every plan shows per-word confidence and flags uncertain words.
Can it read handwriting?
Neat printed handwriting usually reads well. Joined-up cursive is much harder for any OCR engine, including ours — expect errors and proofread the result. We never claim 100% on handwriting; the confidence report tells you honestly how the page went.
What languages are supported?
English by default, plus Spanish, French, German, Portuguese, Italian, Dutch, Chinese (Simplified), Japanese and Arabic, and an English + Spanish mixed mode. Latin-script languages read best; Chinese and Japanese are slower and less accurate, and we say so when you pick them.