Redaction That Can’t Be Undone
Redact PDFs safely and reliably with no residual data. Instead of masking content, Pdftools builds a new PDF without sensitive information. True redaction can’t be reversed and is compliant for heavily regulated industries.
Manual and AI-supported redaction
AI-powered redaction with human-in-the-loop
The AI detection engine identifies sensitive information in a PDF. After review by a user, anything marked for redaction is removed, leaving no residual data behind.
Faster detection with up to 98% accuracy
True redaction with no trace left behind
Append-only audit trail for compliance and troubleshooting
Perform AI-powered redaction with AI Smart Redact
Manual PDF redaction for high-stakes documents
Send a document to the PDF Viewer for manual redaction to completely remove sensitive data—no residual data, no hidden elements.
Highlight elements to mark them for redaction
Sensitive elements are removed, not masked
PDF is rebuilt from scratch, content is not recoverable
Experience manual redaction with the Pdftools PDF Viewer
Read “The Hidden Layers of PDF Redaction” to learn more
Faster and smarter PDF redaction with AI
No genAI, no hallucinations
The detection engine pairs a Named Entity Recognition (NER) model with regular expressions for fine-tuning.
NER extracts information by classifying named entities (people, places, etc.) and is non-generative, which prevents hallucinations and unpredictability; compliance is assured. NER also requires fewer resources and is faster than generative AI. The deterministic part of the engine then fine-tunes detection using regex.
The end user has to review any redaction suggestions before redaction can be applied, so the final decision is always made by a human. The document is then recreated without sensitive data.
Made for secure deployment
Pdftools redaction is self-hosted and deployable in fully air-gapped environments, giving you complete control over data security.
