Free Online OCR — Image to Text
Convert JPG, PNG and WEBP photos, scans or screenshots into editable text. Choose the document language and layout, then review the result before using it.
Upload Image
One image is processed at a time. Original mode sends the selected image pixels straight to the local OCR worker without silent resizing. For a mixed English/Urdu or English/Arabic page, choose the matching combined option.
Extracted Text
Free OCR — Extract Text from Images Online
PDFdukan uses Tesseract.js, a JavaScript/WebAssembly port of the open-source Tesseract OCR engine, to recognise printed text in photos, screenshots and scans. The selector includes 15 individual language models plus English–Urdu and English–Arabic mixed modes. Recognition runs in a browser worker; OCR code and language data are downloaded when needed from configured third-party CDNs.
How to Extract Text from an Image in 3 Steps
- Choose one image — add a JPG, PNG or WEBP photo, screenshot, receipt or scanned page. The tool does not accept PDF files on this page.
- Match the language and layout — choose the primary language or a listed mixed mode. Keep “Automatic page” unless the image is one column, a single block, sparse text or one line.
- Extract, compare and correct — processing time depends on image size, device and whether code/model files are cached. Edit the result before copying or downloading it.
Common Uses for OCR
- Digitising printed documents: Convert photographed letters, certificates, contracts, and reports into editable, searchable text instead of retyping.
- Extracting text from screenshots: Grab text from a screenshot of a webpage, chat, error message, or app where copying is disabled.
- Book and printed notes: Students can extract text from photographed textbook pages or clearly printed notes, subject to copyright and their permitted use.
- Receipts and invoices: Pull amounts, dates, and details from photographed receipts for record-keeping and expense reports.
- Old archives: Convert old printed family documents, certificates, and records into digital text before they fade.
- Translating foreign text: Extract text from an image first, then paste it into a translator — useful for documents, signs, and labels.
Tips for the Best OCR Accuracy
- Use a sharp, well-lit photo — blur is the biggest cause of recognition errors.
- Hold the camera straight above the text — tilted text reduces accuracy. For paper documents, scan with CamMaster Scanner first to flatten and enhance the page.
- Select the correct language before running OCR — the engine loads language-specific recognition models.
- Keep enough detail for readable characters. Tesseract's own guidance says images around 300 DPI often work well; avoid heavily compressed forwards when the original is available.
- Try Document contrast for faded paper or uneven lighting, but compare it with Original mode because preprocessing does not improve every image.
- Use Single column for a simple vertical article and Sparse text for labels or screenshots with separated text.
- Printed text recognises far better than handwriting — Tesseract is built for printed characters.
Frequently Asked Questions
There is no honest fixed accuracy percentage for every image. Clear, upright printed text normally works better than blur, handwriting, tables, decorative fonts or low-contrast scans. Select the correct language and compare names, numbers and other important text with the original.
Yes. Urdu and Arabic are included among the 15 single-language choices. English + Urdu and English + Arabic mixed modes are also available for bilingual pages. The selected model data may download on first use. Review right-to-left text, names, numbers and punctuation against the original.
Tesseract is designed for printed text and performs poorly on cursive or rough handwriting. Very neat, print-style handwriting may partially work. For handwritten notes, expect to correct errors manually.
The OCR workflow processes the selected image in your browser and does not intentionally upload its pixels. The browser does contact third-party CDNs to download Tesseract.js code and selected language data, so those providers receive normal request information such as an IP address. Avoid processing material you are not authorised to handle.
Convert the PDF pages to images first with the PDF to JPG tool, then run OCR on each page image. Alternatively, use the Searchable PDF tool to embed recognised text directly back into the PDF.
On first use, the browser downloads the OCR engine and selected language model data. File sizes vary by model. If the browser keeps those assets in cache, later extractions may start faster.
The listed OCR workflow is currently available without a subscription or account. Practical use is limited by browser memory, language-data downloads and device performance.