PDF OCR — Extract Text from Scanned PDFs in Your Browser

Extract Text from Scanned PDFs in Your Browser

A scanned PDF is essentially a stack of images — the text is locked inside pixels and cannot be selected, searched, or copied. OCR (optical character recognition) reads those pixels, turns them into real text, and produces a searchable PDF you can use in any workflow.

This PDF OCR tool runs entirely in your browser. Files are read with the browser's own APIs and processed on your device. Nothing is uploaded to a server, there is no account, and there is no daily cap beyond a small batch size that keeps the tab responsive.

Why Run OCR Locally

Uploading documents to a third-party OCR service means sending your files somewhere you cannot control. For invoices, contracts, medical records, or any sensitive scan, that is a real privacy concern. Running OCR in your browser keeps the files on your device from start to finish.

Local OCR also means there is no signup, no credit card, and no waiting in a queue. You add the PDFs, click Run OCR, and download the results.

How to Use the Tool

The widget walks through three stages: add PDFs, configure settings, then run OCR and download.

Step 1 — Add PDF files. Click Browse files or drag PDFs onto the drop area. You can add up to three files at once, each up to 20 MB. Only PDF files are accepted. Click Continue to move to the queue.

Step 2 — Configure OCR settings. Each file shows an OCR settings button. Use it to open a modal where you can toggle Fix page rotation (on by default) and set the Document language. Language codes are ISO 639-2 three-letter codes: eng for English, fra for French, deu for German, and so on. Combine languages with commas, for example eng, fra, deu. Click Save settings to apply.

Step 3 — Run OCR. Click Run OCR to process all files one at a time. A progress bar under each card shows which page is being processed. The UI stays responsive — you can see progress without freezing the tab.

Step 4 — Download results. After processing, each card shows Complete or Failed. For successful files, click Save PDF to download the searchable PDF (named {basename}-ocr.pdf) or View text to preview the extracted text in a modal.

Step 5 — Start over. Click Start over to clear everything and return to the intake step.

Examples

English scan. A single-page English contract scanned at 300 DPI. Set language to eng, leave rotation on, run OCR. The result is a searchable PDF with selectable text and a .txt-style preview in the View text modal.

Mixed-language document. A brochure with English and French sections. Set language to eng, fra so Tesseract loads both language models. Accuracy improves noticeably compared to forcing a single language.

Mixed-orientation scan. A document where some pages are portrait and others are landscape. Leave Fix page rotation on so Tesseract detects and corrects the orientation before recognizing text.

Limits

This tool is designed for short, local jobs rather than a full document pipeline:

• Up to 3 files at a time, each up to 20 MB.
• Up to 15 pages per file. Larger documents should be split first.
• Photo quality matters. Clean, high-contrast scans produce the best results. Blurry or low-resolution scans may produce garbled text.
• Handwriting is weak. Tesseract is optimized for printed text. Handwritten notes will have low accuracy.
• No RAW, HEIC, or image-only input. This tool accepts PDF files only. To OCR a single image, use an image conversion tool first.
• Password-protected PDFs cannot be processed and will show an error.

Frequently Asked Questions

Are my PDFs uploaded to a server? No. Files are read in the browser and processed on your device. They are not sent to Quick Tools, stored in localStorage, or attached to a URL.

Which languages are supported? Any language with a Tesseract language model. Common codes include eng (English), fra (French), deu (German), spa (Spanish), ita (Italian), por (Portuguese), rus (Russian), chi_sim (Simplified Chinese), and many more. Language data files are loaded from a public CDN on first use and cached for subsequent runs.

Is the output PDF really searchable? Yes. The output is a standard PDF with an embedded text layer. You can select, copy, and search the text in any PDF reader.

Why did my scan produce poor results? OCR accuracy depends on scan quality. Aim for at least 200 DPI, good contrast, and minimal skew. If the result is poor, try enabling Fix page rotation or improving the source scan.

Can I process more than three PDFs? Not in one batch. Process and download the current batch, click Start over, then add the next set. The limit is there so a single tab stays usable.

Does this tool work offline? The tool itself works offline once the page has loaded. However, Tesseract language data files are fetched from a public CDN on first use for each language. After that, they are cached and available offline.

Related Tools

Looking to shrink images before converting them to PDF? Try the Image Compressor — it re-encodes JPEG, PNG, and WebP at adjustable quality while keeping the original format.