OCR: Extract Text from Images and Scanned PDFs
Choose a photo or a scanned PDF, pick the language and get editable Arabic or English text in seconds, read on your own device.
100% private: your files never leave your browser0 requests to other servers since this page opened.How to verify it yourself
In short
To get text out of a photo or a scanned PDF, choose the file, pick the language of the page and press Extract text. The OCR engine runs inside your browser, reads Arabic and English, and gives you plain text to copy or download. Your file is never uploaded.
A scan or a phone photo of a page is only a picture. You can see the words, but you cannot copy them, search them, translate them or paste them into a form, and job application systems cannot read them either. Optical character recognition, or OCR, looks at the shapes in the image and turns them back into letters you can edit.
Most free OCR websites upload your file to their servers to do this. That is a real risk for the papers people usually need to read: identity cards, certificates, contracts, payslips and medical reports. This tool works differently. It uses Tesseract, the open-source OCR engine first developed at HP and later maintained by Google, compiled to WebAssembly so that it runs on your own device. The engine and its Arabic and English models are downloaded from this site the first time you use it, then your browser keeps them.
Before reading, the tool cleans the image the way a careful scanner would. It brings the page to about 300 DPI, evens out shadows and grey backgrounds so the paper is white everywhere, measures how far the page is tilted and straightens it. This step is what makes a quick photo taken on a desk readable, not only a perfect flatbed scan. Arabic is read right to left with its joined letters, and English words, emails and numbers inside an Arabic line keep their own order.
How to extract text from an image or a scanned PDF
Step 1: Choose the photo or the PDF
Drop a JPG, PNG or WebP photo, or a scanned PDF of up to 50 MB. For a PDF, the first 20 pages are read.
Step 2: Pick the language of the page
Choose Arabic and English for mixed documents, or only the one language on the page: fewer languages read faster and make fewer mistakes.
Step 3: Press Extract text and check it
Each page is cleaned, straightened and read in turn. Pages that were hard to read are named so you can check them against the original.
Step 4: Copy or download the text
Copy everything to your clipboard, or download a UTF-8 .txt file that opens correctly in Notepad, Word and Excel.
Zero-Server
OCR without uploading your documents
The image is decoded, cleaned and read inside a Web Worker in this browser tab. The worker starts only when you press Extract text and is closed a minute after it finishes. No server receives your file, the text or any part of the image. The only downloads are the engine and the language models, from careerpdf.com itself, and they are the same for every visitor.
Nothing about your document is stored: there is no account, no history and no cookie. The privacy meter at the top of the page counts requests to other websites, and it stays at zero while the tool reads your file.
When people use free OCR
- Turning a scanned resume back into text, so you can rebuild it in an ATS-friendly format.
- Copying the details from a certificate, an ID or a residence permit into an online application.
- Getting the text of an old printed letter, contract or report without retyping it.
- Searching and quoting a scanned Arabic book chapter, law or research paper.
- Reading a photo of a whiteboard, a notice or a job advert taken on your phone.
Frequently asked questions
Is this OCR free, and is there a limit?
Yes, it is free with no sign-up and no watermark. A file can be up to 50 MB, and the first 20 pages of a PDF are read. For a longer PDF, split it and read the parts one after another.
How accurate is it for Arabic?
On a clear scan of printed Arabic, almost every line comes out right, with occasional slips in a letter, a dot or the Arabic comma. Quality drops with small or decorative fonts, heavy diacritics, stamps over the text and blurred photos. Handwriting is not supported. Always check names, numbers and dates against the original.
Why does the first use download several megabytes?
The OCR engine and the language models are what actually read the page, and they run on your device instead of a server. Arabic and English together are about 8 MB, downloaded once from this site. After that your browser keeps them and the tool starts at once.
Can I get a searchable PDF instead of plain text?
Not yet. The tool gives you plain text to copy or download. To keep the original layout, paste the text into a document, or rebuild a resume in the resume builder.
My PDF already has text. Should I use OCR?
No. If you can select the words in the PDF, the text is already there and OCR would only add reading errors. The tool tells you when this happens. If the copied Arabic comes out reversed or broken, use the Arabic text fixer instead.
How do I take a photo that reads well?
Lay the page flat, hold the phone straight above it, use daylight or a bright room without a shadow over the text, and make sure the page fills most of the frame. Sharp focus matters more than resolution.
Which browsers does it work in?
Current versions of Chrome, Edge, Firefox, Safari and Samsung Internet on desktop, Android and iPhone. JavaScript and WebAssembly must be enabled, because the page is read on your device.