Someone sends you a WhatsApp screenshot with an IBAN you need to paste into Excel. Or you photographed a six-page contract at the notary and need to search a clause without typing everything by hand. Going from image to text with OCR saves you half an afternoon — if the photo is legible and you know how to review the result.
OCR — optical character recognition — detects letters in the photo and returns editable text. It does not read minds: if the image is blurry or tilted, it invents characters. FORMARTIO extracts in the browser so you can copy, paste, and fix on the fly.
Which photos give good OCR
Printed horizontal text, good light, high contrast — black on white — and no phone shadow across the paragraph. Worse: fast handwriting, tiny decorative fonts, or compressed Telegram screenshots.
Image to text works better if you straighten the page first or crop only the useful zone. Less noise around the edges, fewer confusions for the OCR engine.
Step by step to extract text from an image
- Open Image to Text on FORMARTIO.
- Upload JPG, PNG, or a screenshot.
- Pick the text language if the tool asks — Spanish vs English improves accuracy.
- Run OCR and wait for the result in the text panel.
- Copy, paste into Notes or Word, and fix numbers, IBANs, and proper names by hand.
Mandatory review after OCR
It confuses 0 with O, 1 with l, rn with m. For bank figures and dates, verify character by character against the photo. One wrong digit costs more than typing the whole line.
Photographed tables often come out as messy plain text. Useful to search a word, not to import a perfect Excel sheet. For serious tables, better a direct scan to PDF with a text layer or a table-specific tool.
Image to text in practical cases
Business cards at events: OCR the contact and paste into your CRM. Google Lens translations compete, but sometimes you want clean text without an extra app.
Whiteboard notes in a meeting: photo + image to text + summary in Notion. The original photo stays as proof if OCR fails on a technical acronym.
Paper invoices when the vendor does not send PDF: extract tax ID, total, and date for your expense sheet. Double-check VAT and decimals.
How to improve results without pro gear
Shoot perpendicular, fill the frame with the document, avoid direct flash that blows out the center. Two photos with different light: keep the one OCR reads better.
Gentle contrast boost in an editor before OCR can help on grayish photos. Do not go to extreme black and white or you lose fine serifs.
Handwriting: OCR improves every year, but still fails on doctor scribble. For short personal notes, try it; for legal, human transcription.
Mixed Spanish-English on the same photo — imported product labels —: try OCR in the dominant language and review technical terms in the other language by hand.
Image to text and privacy
Do not upload screenshots with passwords, tokens, or medical data to unknown services. FORMARTIO processes in the browser; still, delete sensitive captures from disk after you copy what you need.
For long scanned PDFs, OCR on loose pages sometimes works better than a photo of a folded book with spine shadow.
When another unreadable screenshot lands and you are about to type it by hand, try automatic extraction first. Extract text from your image on FORMARTIO, review critical figures, and paste where you need it.