Hey Product Hunt!

I'm Leo, founder of Document Redactor.

This project started when I watched a friend spend hours each week manually redacting PDFs for compliance responses — and then paste the file into ChatGPT to "speed it up." The irony of trying to protect personal data by handing it to OpenAI was too perfect to ignore.

Most AI redaction tools today stream your documents to third-party LLM APIs (OpenAI, Anthropic, Google). That's exactly the data-transfer problem you're trying to solve in the first place.

So we built Document Redactor differently:
- Detection runs on local AI models on our own servers, or optionally on EU-hosted Mistral. Never OpenAI, Anthropic, or Google.
- OCR built in — scanned PDFs are auto-rotated and OCR'd before detection runs
- Review before redact — see every match on the page, toggle hits, filter by category, draw manual boxes
- Files auto-deleted 2 hours after processing — no silent retention, no shadow archives
- Batch processing — drop multiple files, get one ZIP back
- 6 languages: EN, ES, FR, DE, PT, IT
- Real redaction — for text PDFs, we strip the underlying characters and metadata, not just paint a black box on top

You can preview detections for free, no account needed. You only pay when you download a redacted file.

Plans start around $8/month yearly billing, up to 10,000 docs/month at Enterprise Plus.

I'd love your feedback - especially:
1. What's the worst redaction pain you've hit?
2. If you work in a regulated field (legal, healthcare, HR, gov) – what would make you switch tools?
3. Any entity type you wish AI redaction caught that nothing does today?

🎁 Product Hunt offer: use code PRODUCTHUNT (10 credits for free).

Document Redactor

AI PDF redaction that never sends your docs to public LLMs

AI PDF redaction that never sends your docs to public LLMs