Image to Word: Free OCR in English, French, Arabic
To convert an image to editable text for free, upload the photo or scan to myoffice, open the file's menu, and choose Convert to Word (OCR) for a .docx or Convert to PDF (OCR) for a searchable PDF. Recognition supports English, French and Arabic. You need a free account, and the result is saved as a new file next to the original.
This guide explains what OCR is, which output to choose, and the handful of habits that make the difference between clean text and a mess.
What is OCR?
OCR (optical character recognition) reads the letters in an image and converts them into real text that you can select, search, copy and edit. A photo of a page is just pixels. After OCR, the same page contains actual characters.
myoffice uses Tesseract, a widely used open-source OCR engine, running on the server. Your image is processed as part of your account's storage and is not sent to a third-party OCR service.
How do I convert an image to Word?
- Sign in or create a free account.
- Upload the image (JPG, JPEG, PNG, BMP, TIFF or WebP, up to 50 MB).
- Open the file's ⋮ menu (or right-click it).
- Choose Convert to Word (OCR).
- Wait for the job to finish in the job queue. A new
.docxappears in the same folder. - Open it in the browser editor to correct anything and format it.
Word or searchable PDF: which should I pick?
| Convert to Word (OCR) | Convert to PDF (OCR) | |
|---|---|---|
| Output | A .docx with the recognised text, one paragraph per line | The original image with an invisible text layer on top |
| Best for | Editing and reusing the text | Archiving a document that must keep its original look |
| Layout | Plain text, so you will reformat | Identical to the scan |
| Searchable | Yes | Yes, and you can select and copy text |
Choose Word when you want to change the words. Choose searchable PDF when you want to keep the original appearance, such as a signed page or an official form, but make it findable and copyable.
Which languages does the OCR support?
English, French and Arabic. The recogniser is configured for all three together, so a page that mixes them, such as a French document with an Arabic name or an English form with an Arabic address, is read in a single pass without you picking a language first.
Two caveats worth knowing:
- Handwriting is not reliably recognised. OCR is designed for printed text.
- Arabic is the hardest of the three, because letters change shape depending on position and small dots separate similar letters. Clean, high-resolution input matters most here.
How do I get accurate OCR results?
Accuracy depends far more on the image than on the tool. These habits help with any OCR engine:
- Aim for about 300 DPI. For a scan, choose 300 DPI. For a phone photo, get close and fill the frame with the page.
- Keep the page flat and straight. Skew, curved page edges and perspective distortion all cost accuracy.
- Use even light and good contrast. Dark text on a plain light background is ideal. Avoid shadows across the page.
- Avoid heavy compression. A blocky, over-compressed JPG loses the small details, especially Arabic dots and French accents.
- Use a clean original. A photocopy of a photocopy reads worse than the first-generation page.
- Proofread the result. Even good OCR makes mistakes on numbers, names and similar characters such as
Oand0. Check anything that matters.
What if my document is a multi-page PDF?
The OCR tool works on image files. For a stack of pages photographed one by one, the easiest route is to run OCR on each image, or to combine the images with Images to PDF and use the PDF tools to organise the pages.
Is it private?
Uploads are scanned for malware before they are stored, and the OCR runs on the same server as your files rather than a separate outside service. Your document contents are not sold or mined. For very sensitive material, delete the source and the output from your storage once you are finished, and protect any PDF you keep with a password.
Frequently asked questions
Can OCR read Arabic text?
Yes. Arabic is supported alongside English and French. For best results, scan at 300 DPI or higher and use a clean, high-contrast original.
Is the OCR free?
Yes. It is part of the free plan. Files count against your storage (100 to 250 MB on the free plan).
Does the Word file keep my layout?
No. The Word output contains the recognised text as paragraphs, so tables, columns and images are not reproduced. To keep the exact layout, choose the searchable PDF output instead.
What image formats can I use?
JPG, JPEG, PNG, BMP, TIFF and WebP.
Try it
Create a free account, upload a photo of a printed page, and choose Convert to Word (OCR). Then compare the output with the original to see how well your image quality worked.