Image to Text (OCR): Extract Text from an Image Privately, or Clean Up a Photo So Any OCR App Can Read It
A screenshot of an error message, a photo of a receipt, a handout you do not want to retype: each is an image to text job, and the technology behind it is called OCR, optical character recognition. This guide shows how to extract text from an image with the free image to text converter above, why the same photo reads well in one app and returns garbage in another, and how to clean up a picture so the text comes out right — privately, with no sign-up.
1. What Image to Text (OCR) Is, and When You Need It
A picture of words is still a picture: you can read it, but you cannot select, search or edit it. OCR decides which shapes are letters and hands you real characters. Whether it is a JPG to text job on a phone snap or a PNG to text job on a screenshot, the goal is text you can work with.
Use it whenever retyping would be slow or error-prone. What few guides explain is that recognition quality depends more on the image than on the software: small, faint or light-on-dark text trips up every OCR tool, while a cleaned-up copy of the same picture reads well. That is why the tool above is built around preparation first.
2. How to Extract Text from an Image in Four Steps
Here is how to extract text from an image with the tool above, one image at a time, with no account or daily limit.
- Add the image. Drop, choose or simply paste it: a screenshot on your clipboard goes straight in, the quickest screenshot to text route there is. Very large images are scaled so their longest side is at most 2,500 pixels.
- Check the preview. By default the image is enlarged 2× and converted to pure black and white with an automatic threshold. The preview updates as you change settings.
- Adjust if needed. Tick Invert for light text on dark, try 3× for tiny print, or move the threshold if strokes break up or fill in.
- Get the result. Where the panel says text recognition is available, press Recognise text, then copy the text or download it as a file ending in -ocr.txt. Where it is not available, the button reads Download prepared image and gives you a clean PNG ending in -ocr-ready.png for the apps in section 5.
A straight answer on that second case: built-in text recognition is currently offered here only in Chrome or Edge with the experimental web platform features setting switched on, so most visitors get the prepared image. The panel says which you have before you press anything, and your picture is never passed to an outside OCR service or stored on our servers.
3. How to Prepare an Image for OCR
If you have wondered how to prepare an image for OCR or how to improve OCR accuracy, these four adjustments matter most, and each is a control in the panel above.
- Upscale. Recognition needs enough pixels per letter. If you need to upscale small text image for OCR work, 2× is the default and 3× helps the smallest print. It cannot invent detail the camera never caught.
- Grayscale. Removes tinted paper, coloured highlights and colour casts that confuse recognition.
- Binarize and threshold. The step people search for as binarize image for OCR forces every pixel to pure black or pure white. With the threshold slider at “auto”, the cut-off is chosen for you from the image’s own light and dark tones using Otsu’s method. Set your own value from 1 to 255 instead: lower keeps only the darkest ink, higher keeps faint strokes. It is the most direct way to increase image contrast for OCR.
- Invert. Many readers expect dark text on a light page. Ticking invert flips a dark-mode screenshot or a white-on-black slide into that form.
Note that once binarize or invert is on, the image is turned grey anyway, and the threshold only matters while binarize is ticked.
4. OCR Not Recognizing Text? The Fixes That Work
When an app returns nothing or a string of symbols, match the symptom to the fix.
- OCR not recognizing text at all: Usually the text is too small or too faint. Upscale to 3×, keep binarize on with the threshold at auto, and check that letters are solid black shapes with clean gaps.
- OCR white text on black background: Dark-mode screenshots and slides defeat many readers. Tick Invert so the prepared image shows black text on white.
- Blurry image to text: Upscale 2× and nudge the threshold lower until letters separate. If the preview is still a smudge, retake the photo.
- Faint print, such as an old receipt: Raise the threshold so light grey strokes count as ink.
- Phone camera text tools not detecting text: The text features built into phones struggle with tiny lettering, glare and busy backgrounds too. Open the prepared image in them instead of the original photo.
- Crooked, sideways or cluttered photos: This tool does not crop or rotate. Trim to the text area with the image cropper and turn it upright with rotate and flip before you bring it back here.
5. Copy Text from a Picture on Windows, Mac, iPhone and Chrome
If the panel offers you the prepared image, the text tools already built into your system finish the job. Open the -ocr-ready.png in whichever of these you have.
- How to copy text from a picture on Windows: Open the image full-screen, capture it with Snipping Tool on Windows 11 and use Text actions to copy the words. If you use Microsoft’s PowerToys, its Text Extractor lets you draw a box around any text on screen and copies it straight to the clipboard.
- How to copy text from an image on Mac: Open the file in Preview; on recent versions of macOS, Live Text lets you drag across the words and copy them like ordinary text.
- Image to text for iPhone photos: Live Text works in the Photos app on recent iPhones: press and hold on the words. If an iPhone photo will not open here, it is probably HEIC; convert it with HEIC to JPG first, and use AVIF to JPG the same way for AVIF files.
- Extract text from a screenshot in Chrome: Open the prepared image in a tab, right-click it, choose Search image with Google and select the text.
6. Receipts, Invoices, Whiteboards, and What This Tool Does Not Do
A receipt photo to text job is a faint-print problem: raise the threshold and upscale. For invoice image to text work, privacy matters as much as accuracy, and your file is never stored on our servers. A whiteboard photo to text job needs the glare gone: crop tightly and lower the threshold so reflections turn white. For scanned image to text for students, a textbook page or handout usually reads well at the defaults; paste the result into the word counter or tidy capitals with the case converter.
Just as important is what this tool does not do:
- Scanned PDFs. It takes images only. To make a scanned PDF searchable, use PDF OCR; to get photos into a searchable PDF, combine them with image to PDF first.
- Tables and layout. The output is plain text with no columns or formatting, and there is no photo-to-Excel path. For PDFs, the PDF to Excel and PDF to Word converters keep structure.
- Batches, languages and handwriting. One image at a time, no language setting, and handwriting is not read reliably.
- Translation, QR codes and barcodes. Not supported; our QR generator only creates codes.
- Hiding text. To obscure words in a picture, use the image blur tool, or PDF redact for documents.
7. Why Choose Cortexis as Your Free Image to Text Converter
Plenty of pages advertise a photo to text converter or free online OCR, but few say what happens to your picture or let you fix a bad one. Here is why this image to text converter online free is different.
Genuinely free, no account
No sign-up, no email, no captcha and no daily file counter.
Fixes the image, not just the text
Upscale, grayscale, automatic black-and-white threshold and invert, with a live preview.
Honest about what you get
You are told up front whether you will get copyable text or a prepared image. No accuracy promises.
Your images stay private
Screenshots, invoices and IDs are never stored on our servers and never retained, and never shared with an outside OCR service.
Nothing is used for training
Your images are never used to train any model.
Works with the tools you already have
The prepared PNG is made to be opened in Snipping Tool, Preview and Live Text, with clean black text on white.
It sits among a wider set of free file tools built on the same terms.
8. Frequently Asked Questions (FAQ)
How do I convert an image to text for free?
Drop, choose or paste the image into the tool above and check the preview. Where text recognition is available, press Recognise text and copy or download the result. Otherwise download the prepared image and open it in Snipping Tool, Preview or Live Text.
Why do I get a prepared image instead of text?
Built-in text recognition is currently offered here only in Chrome or Edge with the experimental web platform features setting switched on. Everywhere else you get a cleaned, enlarged black-and-white PNG, which the text tools in Windows, macOS and iPhone read more reliably.
Is online OCR safe for confidential documents?
It depends on the site, and many do not say what happens to your file. Here, images are never stored on our servers, never retained and never used for training.
Can I extract text from image files in bulk?
No, it works on one image at a time, and none of our other tools batch image OCR either. For a stack of scanned pages, combine them with image to PDF and run PDF OCR on that one file.
Can it read text from a scanned PDF?
No, it accepts images only. Use the PDF OCR tool to make a scanned PDF searchable, or combine photos with image to PDF first and then run PDF OCR on the result.
Can I convert a photo of a table into Excel?
Not with this tool; the output is plain text without columns. If the table is in a PDF, PDF to Excel keeps rows and columns.
Does it read handwriting or other languages?
There is no language setting, and handwriting is not read reliably. Clear printed text, such as typed pages, screenshots and receipts, gives the best results.
My iPhone photo will not open. What should I do?
iPhone photos are often saved as HEIC, which some setups cannot open. Convert it with HEIC to JPG, then drop the JPG here; AVIF to JPG does the same for AVIF files.
Need the picture for something else? Remove location data with the EXIF remover, enlarge a low-resolution photo with the image enlarger, or shrink it for email with the image compressor. If you would like to learn to program, our free Python development course is a good start.