How to convert an image to text
- Add your images. Drop JPG, PNG, WebP, BMP or GIF files onto the box, tap to choose them, or copy a screenshot and paste it with Ctrl+V (⌘+V on a Mac).
- Choose the text language. Tap each language that appears in the images: English, Urdu, Spanish, Portuguese or Bahasa Indonesia. Pick more than one for mixed text.
- Press Extract text. A progress bar shows the engine loading and each image being read. You can press Stop at any time.
- Check and correct the result. The text appears in an editable box. Fix any misread words directly there.
- Copy or download. Use Copy to put the text on your clipboard, or Download .txt to save it as a plain text file.
Features
- Five languages, matching the languages of this site, and any combination of them for mixed-language images.
- Batch reading: add several images and get all their text in one place, each under its file name.
- Paste from the clipboard: grab a screenshot and paste it straight in, no saving needed.
- Editable result with one-tap copy and a plain text download.
- Right-to-left display for Urdu, so extracted Urdu text reads the right way round.
- Automatic image preparation: small images such as cropped screenshots are enlarged before reading, which helps with small text, and transparent PNGs are placed on white so dark text doesn't vanish.
- Stop button for long batches, keeping any text already read.
Why use Fileora to extract text from images
Most online OCR services upload your picture to a server and read it there. That's a poor fit for the things people most often need to copy text from: ID cards, bank letters, receipts, medical reports and private messages. Fileora does the recognition in your browser with an open-source engine (Tesseract), so the picture stays on your phone or computer. There's no account, no daily limit and no watermark.
Because nothing is uploaded, a large batch doesn't have to travel over a slow connection either. Once the engine and language files are stored by your browser, reading works quickly even on mobile data.
Getting the best results
The quality of the photo matters more than anything else. Hold the camera straight above the page so the lines of text are level, fill the frame with the text you need, and make sure the light is even with no shadow from your hand or phone. Blurry, tilted or low-contrast images are the most common cause of mistakes.
If you only need one paragraph from a busy page, crop it first with crop image. Less clutter means fewer stray characters. Screenshots usually read almost perfectly because the text is sharp and perfectly straight.
Only tick the languages that are really in the image. Each extra language slows reading down a little and can make the engine guess between similar-looking letters. For English text with a few Spanish names, English alone is usually enough; for a Pakistani form with both Urdu and English, tick both.
iPhone photos saved as HEIC may not open in every browser. Convert them first with HEIC to JPG, then add the JPG files here.
What OCR can and can't do
Optical character recognition turns the shapes of letters into editable text. It is excellent at clean printed text and gets worse as the image moves away from that. Tables come out as plain lines of text, not as a spreadsheet. Columns in newspapers and leaflets may be read in an unexpected order. Logos, stylised headings and text on photos are hit and miss. Handwriting is largely beyond it.
Urdu deserves a special note. The Urdu language data is trained mainly on Naskh-style letters, so on-screen Urdu in common web fonts reads well, while traditional Nastaliq type from books and newspapers is often split or joined incorrectly. Treat Nastaliq results as a rough draft that needs careful checking.
Once you have your text, you can count its words and characters with the word counter. For documents that are already PDFs, PDF to Text is faster and exact when the PDF contains real text rather than scanned pictures.