Pragma Redact
Automated PDF, text and image anonymization - AI and smart patterns, 100% local
Pragma Redact finds and masks sensitive information in PDF documents, text and images on its own. An AI model understands sentence context and recognises names or addresses, while smart pattern rules instantly catch formatted identifiers: bank accounts, tax and ID numbers, e-mail addresses. No more redacting documents by hand. The app runs in your browser, with no sign-up and no quota.
Interface language: English, Polish
Open the tool
What it does
Automatic detection of personal data by an AI model that understands context
Smart pattern rules for account numbers, identifiers, e-mail addresses and contact details
Anonymization of PDF files, plain text and images (with OCR text recognition)
Uses your device's graphics card to speed things up when one is available
The model is downloaded once and cached in the browser - later visits are much faster
Manual adjustment of the masks before saving the result
Interface available in English and Polish
When to use it
- • Clean up prompts before pasting a document into ChatGPT, Claude or any other chatbot
- • Strip customer data from support tickets, logs and reports before sharing them further
- • Safely share screenshots and document snippets outside your organisation
- • Catch API keys and tokens accidentally left in documentation or code
- • Prepare invoices and contracts for an audit, a portfolio or training material
Privacy
No file is ever sent to the internet. The entire application - including the AI model - runs in your browser, on your device.
Technologies and competences
Need data anonymization in your own processes?
Pragma Redact shows how far local, on-device AI can go without sending data anywhere. We can bring the same approach into your organisation.
Talk to us about AI adoption