OCR Image
Open the app: Turn an image into text → (products.documentize.app)
In one sentence
OCR Image reads the text in your pictures (photos, screenshots, scans saved as images) and hands it back as plain .txt files.
What makes it unique
- It starts from images, not PDFs, and it ends with plain text, not a PDF. That is the opposite of Searchable (PDF in, PDF out) and different from Extract Text (PDF in, reads text that already exists).
- It is a batch tool: you can drop many images at once and get one text file per image, or tick one box to receive everything in a single text file.
- You pick the document language from a list of about 35 languages (English, Arabic, Chinese simplified and traditional, Cyrillic languages, Hindi, Japanese, Korean, Thai and more).
The site's FAQ also says it copes with printed text, screenshots and photographs, and "even handwritten text". Treat the handwriting claim with caution: results on handwriting vary a lot (verify on the live app with your own samples).
When to use it / when not to
- Use it when you have a photo of a page, a receipt, a whiteboard, a screenshot or a scanned image and you only want the words, to paste into an email or an editor.
- Do not use it when you need to keep the page looking the same and also be able to search it. Use Searchable: it keeps the picture and hides the text underneath.
- Do not use it on a PDF that already has selectable text; Extract Text is faster and exact.
- Do not use it if you want a Word, Excel or PowerPoint file. Searchable can output those.
Before you start
- Have your images as JPG or PNG (the web page accepts these two; the engine behind it also handles TIFF, BMP and some other formats, and other formats are converted automatically, but upload JPG/PNG to be safe).
- Make sure the photo is upright. Sideways pages read badly; turn them first (see Rotation for PDFs, or rotate the picture on your device).
- Know the language of the text. Choosing the wrong language is the number one cause of garbage output.
- The site's FAQ states up to 100 files and about 32 MB in total per process.
Step by step
- Open OCR Image (the page is titled "Image to TXT") and drop your pictures.
- Open Show all settings and choose the Document language.
- Optionally tick Convert to single output file if you want all texts merged into one .txt.
- Press Recognize.
- Download the result. Press Recognize again to start another job.
📷 Screenshot placeholder: the OCR Image page with the language list open
Options explained
| Option | What it does | Default | Recommended |
|---|---|---|---|
| Document language | Tells the engine which alphabet and words to expect | English | Always set it to the language of the text in the picture |
| Convert to single output file | Joins the text of all images into one .txt (only matters with more than one image) | off | On when the images are pages of one document, in the right order; off when they are unrelated |
| Landscape orientation | Appears on the page, but the text recognition itself does not use it (verify on the live app) | off | Leave it off; rotate the image beforehand instead |
What you get
- One image gives one .txt file with the same base name.
- Several images give a ZIP called
images_to_txtwith one .txt per image, unless you chose the single-file option, in which case you get one text file. - Your original images are not modified.
- If one image cannot be read at all, the others are still processed; if none can, the job reports an error.
Quality & limits
- Recognized, not copied. This is an estimate of what the picture says, so expect some wrong letters, especially with blur, glare, shadows, small print, tables or decorative fonts.
- Plain text has no layout: columns, tables and fonts are lost.
- Because each image is read on its own, a free-plan time limit can stop a big batch early. You may then be offered a "continuation" to finish the rest later (a cool-down of about 12 hours applies by default). With the single-file option the batch is treated as one piece, so it cannot be continued that way.
- Results are kept on the server for about two days.
Works well with
- Use before: Rotation, Deskew for crooked scans, Resize is rarely needed.
- Use after: paste the text into your editor, or build a PDF from the images with Merger and make it searchable with Searchable if you still need the pages.
- Recipes: Photos to searchable PDF, Digitise an archive.
Common mistakes & fixes
- "The text is gibberish" → wrong language, or the picture is sideways or too small. Fix both and try again.
- "I wanted a PDF" → this app returns text. Use Searchable (convert images to PDF first with Merger).
- "The order of pages in the single file is wrong" → the order follows your upload order; re-upload in order.
- "Tables came out as lines of words" → plain text cannot keep tables; try Searchable with Excel output and "tables only".
Same thing elsewhere
- CLI plugin: Documentize.OcrImage (needs the Tesseract program installed on your computer; it reads images from an input folder and writes text to an output folder).
- MCP tool: RunOcr (its description mentions a PDF result, while the web app returns text; verify which you get).
- See Beyond the browser.
Try it
Take a photo of a printed page (a menu or a letter) with your phone, upload it, choose the correct language and press Recognize. Open the .txt and count the mistakes. Then take a second photo at an angle in poor light and compare: you will see why a straight, well-lit photo matters.