Start here

Quick start & APIInstallOnline OCRCode generator

Models & pipelines

PP-OCRv6 modelsLanguagesPP-StructureV3PaddleOCR-VL

Deploy & integrate

Serving & APIPaddleOCR.jsMCP server

Fix & upgrade

TroubleshootingMigrate from 2.xReleases

Evaluate

Our benchmarkCompare OCR toolsLicense & cost

Tools

Online OCRInstall command builderCode generatorError fixer2.x → 3.x converterLanguage code finder

Language

English简体中文日本語Try OCR online

Privacy policy

Short version: we use Google Analytics to see which pages help people. The online OCR tool runs in your browser and never uploads your images. There are no accounts, forms or ads.

Updated

No uploads#

The online OCR tool processes images on your device. We never receive your images or the extracted text.

Analytics only#

Google Analytics 4 measures visits. That's the only tracking on the site.

No ads today#

No advertising scripts are loaded. This page will change before any are added.

What is collected, and by whom#

  • Google Analytics 4 (measurement ID G-2Y98KKZZWL) runs on every page. It sets cookies and records page views, approximate location derived from your IP address, device and browser type, referrer, and a few anonymous tool events (for example "OCR run finished, 27 lines, 2 seconds" — never the text or image). See how Google uses data from sites that use its services. You can opt out with the Google Analytics opt-out add-on or by blocking analytics in your browser.
  • Google Fonts (Geist and Geist Mono) load from fonts.googleapis.com and fonts.gstatic.com, so Google receives your IP address and browser details when a page loads.
  • Hosting: the site is served by Vercel, which keeps standard server logs (IP address, user agent, URL, time) for operations and security.

The online OCR tool#

When you use Online OCR, your image is read into the page and processed by ONNX Runtime Web inside your browser tab. It is not uploaded to paddleocr.dev or anyone else. To do that, your browser downloads:

  • the PaddleOCR.js SDK code from jsDelivr (cdn.jsdelivr.net);
  • model files from Baidu's CDN (paddle-model-ecology.bj.bcebos.com) or, if that fails, from Hugging Face (huggingface.co).

Those services see a normal download request (your IP address and browser), not your image. The models are then kept in your browser's Cache Storage so later visits don't download them again; you can remove them by clearing this site's data in your browser settings.

What stays in your browser#

  • paddleocr_theme (localStorage) — your light/dark choice.
  • pocr_tab_* (localStorage) — which code tab you picked (for example "NVIDIA GPU"), so other examples match.
  • pocr-models-v1 (Cache Storage) — downloaded OCR models, used only by the online tool.

None of these are sent to us.

Other tools#

The install builder, code generator, error fixer, migration converter and language finder run entirely in your browser on what you type or select. Text you paste into them (for example a traceback or old code) is not sent anywhere.

PaddleOCR itself#

The PaddleOCR library runs on your own computer or server, so files you process with it never reach paddleocr.dev. It downloads model files from a model hub the first time you use a model. Some optional modes — the hosted PaddleOCR API, or the MCP server's AI Studio and Qianfan modes — send your files to Baidu's services; check the terms of whichever mode you use.

Your choices and rights#

Depending on where you live (for example under the GDPR or UK GDPR) you may have rights to access, correct or delete personal data and to object to its processing. We don't hold personal data ourselves beyond what the providers above process for us; for analytics data you can also use Google's own controls such as Google Activity controls.

Children#

The site is written for developers and isn't directed at children.

Contact and changes#

Questions about privacy: hello@paddleocr.dev. If this policy changes, the date at the top of the page changes with it.