# Safe-Doc.ai

> Safe-Doc.ai is the document-security layer that pseudonymizes your documents before any AI processing (ChatGPT, Claude, Gemini, Copilot). Pseudonymization runs on self-hosted models in the European Union: no third-party AI sees your content, and nothing is used for training. Real values are restored locally on the client side. Access via web app, API and MCP server. Hosted in the EU (Hetzner), ephemeral processing, zero storage.

Tagline: pseudonymization before AI · your documents, safe in AI. Canonical domain: https://safe-doc.ai/ · Contact: contact@safe-doc.ai · French index: https://safe-doc.ai/llms-fr.txt

**Access modes:** web UI, REST API, MCP server (open standard) — no copy-paste required for API/MCP workflows. API documentation (OpenAPI): https://app.safe-doc.ai/docs · MCP server package: @safedoc/mcp

**Differentiators:** 55+ entity types, multinational coverage (FR, DE, ES, IT, UK, US); zero storage after processing; no persistent workspace; Data Room for consistent pseudonyms across batches; reversible pseudonymization via local mapping JSON; REST API + MCP (Cabinet+, Enterprise); model-agnostic layer, not a closed platform.

**Product modes:** Clean (daily use, metadata strip, pseudonymize or anonymize) and Advanced (high-sensitivity cases, Data Room).

**Hosting & security:** Self-hosted models in the EU; no third-party AI on document content. TLS 1.2+ / 1.3; ephemeral session data (2 h purge); stateless mode available. EU hosting (Hetzner, Germany). Subprocessors: Hetzner (processing), Scaleway (email), Zoho (team email), Namecheap (DNS), Stripe (billing only — never your documents). Host certifications: ISO/IEC 27001:2022 and BSI C5 Type 2 (Hetzner, not Safe-Doc.ai).

**Pricing (excl. tax):** 14-day free trial, no credit card (Solo, 1 seat). Solo €35/month · Cabinet €149/month · Cabinet+ €399/month · Enterprise custom.

## Core pages

- [Home (EN)](https://safe-doc.ai/index_us.html): what Safe-Doc.ai does and who it is for
- [How it works](https://safe-doc.ai/how.html): pseudonymize, use AI, then restore the output
- [Formats & PII coverage](https://safe-doc.ai/formats.html): supported file types and detected entities
- [FAQ](https://safe-doc.ai/faq.html): security, GDPR and product usage
- [Shadow AI (risks)](https://safe-doc.ai/shadow-ai.html): uncontrolled AI use at work, with sourced figures
- [Pseudonymization vs anonymization](https://safe-doc.ai/pseudonymisation.html): two regimes and when to use each
- [Why a layer](https://safe-doc.ai/why-a-layer.html): model-agnostic protection layer vs closed platforms
- [Security & architecture](https://safe-doc.ai/security.html): hosting, subprocessors, logs, zero storage
- [Integrations (API + MCP)](https://safe-doc.ai/integrations.html): REST API and MCP server access
- [Pricing](https://safe-doc.ai/pricing.html): plans, seats and prices
- [14-day free trial](https://safe-doc.ai/signup.html): start Solo trial without credit card
- [Demo](https://safe-doc.ai/demo.html): interactive demo without account
- [About](https://safe-doc.ai/about.html): company and product context
- [Glossary](https://safe-doc.ai/glossary.html): key terms defined
- [Contact](https://safe-doc.ai/contact.html): sales and support contact
- [Legal notice](https://safe-doc.ai/mentions-legales.html): publisher and legal information
- [DPA](https://safe-doc.ai/dpa.html): data processing agreement

## Industry segments

- [Law firms](https://safe-doc.ai/law-firms.html): case files and professional secrecy before AI
- [Accounting](https://safe-doc.ai/accounting.html): financial and audit documents before AI
- [HR](https://safe-doc.ai/hr.html): employee and HR data before AI
- [M&A & due diligence](https://safe-doc.ai/ma-due-diligence.html): deal documents and data rooms
- [Consulting](https://safe-doc.ai/conseil.html): client confidential material before AI
- [Legal departments](https://safe-doc.ai/direction-juridique.html): in-house legal teams
- [DPO landing](https://safe-doc.ai/dpo.html): data protection officers and privacy teams
- [Security teams](https://safe-doc.ai/security-teams.html): CISO and security teams

## Intent landings (SEO)

- [Anonymize a PDF (FR)](https://safe-doc.ai/anonymiser-un-pdf.html): French landing — pseudonymize a PDF before AI
- [Anonymize a Word document (FR)](https://safe-doc.ai/anonymiser-un-document-word.html): French landing — .docx before AI
- [Secure ChatGPT for enterprise (FR)](https://safe-doc.ai/securiser-chatgpt-entreprise.html): French landing — controlled AI at work
- [GDPR pseudonymization (FR)](https://safe-doc.ai/pseudonymisation-rgpd.html): French landing — GDPR pseudonymization
- [Anonymize before ChatGPT (FR)](https://safe-doc.ai/anonymiser-un-document-avant-chatgpt.html): French landing — prepare a document

## Blog

- [Blog hub (EN)](https://safe-doc.ai/blog/index.html): articles on AI confidentiality and GDPR
- [Shadow AI: the numbers](https://safe-doc.ai/blog/shadow-ai-by-the-numbers.html): sourced data on shadow AI
- [Pseudonymization vs anonymization](https://safe-doc.ai/blog/pseudonymization-vs-anonymization.html): GDPR art. 4(5) explained
- [Lawyers & professional secrecy](https://safe-doc.ai/blog/lawyers-chatgpt-professional-secrecy.html): ChatGPT under professional secrecy
- [ChatGPT & confidential documents](https://safe-doc.ai/blog/chatgpt-confidential-documents.html): risks and safe practices

## Optional

- [Advanced mode](https://safe-doc.ai/advanced.html): high-sensitivity processing and Data Room
- [Beta program](https://safe-doc.ai/beta.html): beta access information
- [Terms of Use](https://safe-doc.ai/cgu.html): terms of use
- [Terms of Sale](https://safe-doc.ai/cgv.html): terms of sale
- [Legal](https://safe-doc.ai/legal.html): legal hub
- [Integrations (FR)](https://safe-doc.ai/integrations-fr.html): French integrations page
- [Cowork](https://safe-doc.ai/cowork.html): Claude Cowork integration (gated preview)
- [About (Markdown)](https://safe-doc.ai/about.md): Markdown mirror
- [Accounting (Markdown)](https://safe-doc.ai/accounting.md): Markdown mirror
- [Advanced (Markdown)](https://safe-doc.ai/advanced.md): Markdown mirror
- [Beta (Markdown)](https://safe-doc.ai/beta.md): Markdown mirror
- [Chatgpt Confidential Documents (Markdown)](https://safe-doc.ai/blog/chatgpt-confidential-documents.md): Markdown mirror
- [Lawyers Chatgpt Professional Secrecy (Markdown)](https://safe-doc.ai/blog/lawyers-chatgpt-professional-secrecy.md): Markdown mirror
- [Pseudonymization Vs Anonymization (Markdown)](https://safe-doc.ai/blog/pseudonymization-vs-anonymization.md): Markdown mirror
- [Shadow Ai By The Numbers (Markdown)](https://safe-doc.ai/blog/shadow-ai-by-the-numbers.md): Markdown mirror
- [Cgu (Markdown)](https://safe-doc.ai/cgu.md): Markdown mirror
- [Cgv (Markdown)](https://safe-doc.ai/cgv.md): Markdown mirror
- [Conseil (Markdown)](https://safe-doc.ai/conseil.md): Markdown mirror
- [Contact (Markdown)](https://safe-doc.ai/contact.md): Markdown mirror
- [Demo (Markdown)](https://safe-doc.ai/demo.md): Markdown mirror
- [Direction Juridique (Markdown)](https://safe-doc.ai/direction-juridique.md): Markdown mirror
- [Dpa (Markdown)](https://safe-doc.ai/dpa.md): Markdown mirror
- [Dpo (Markdown)](https://safe-doc.ai/dpo.md): Markdown mirror
- [Faq (Markdown)](https://safe-doc.ai/faq.md): Markdown mirror
- [Formats (Markdown)](https://safe-doc.ai/formats.md): Markdown mirror
- [Glossary (Markdown)](https://safe-doc.ai/glossary.md): Markdown mirror
- [How (Markdown)](https://safe-doc.ai/how.md): Markdown mirror
- [Hr (Markdown)](https://safe-doc.ai/hr.md): Markdown mirror
- [Home (Markdown)](https://safe-doc.ai/index_us.md): Markdown mirror
- [Integrations (Markdown)](https://safe-doc.ai/integrations.md): Markdown mirror
- [Law Firms (Markdown)](https://safe-doc.ai/law-firms.md): Markdown mirror
- [Legal (Markdown)](https://safe-doc.ai/legal.md): Markdown mirror
- [Ma Due Diligence (Markdown)](https://safe-doc.ai/ma-due-diligence.md): Markdown mirror
- [Mentions Legales (Markdown)](https://safe-doc.ai/mentions-legales.md): Markdown mirror
- [Footer En (Markdown)](https://safe-doc.ai/partials/footer-en.md): Markdown mirror
- [Gtag (Markdown)](https://safe-doc.ai/partials/gtag.md): Markdown mirror
- [Pricing (Markdown)](https://safe-doc.ai/pricing.md): Markdown mirror
- [Pseudonymisation (Markdown)](https://safe-doc.ai/pseudonymisation.md): Markdown mirror
- [Security Teams (Markdown)](https://safe-doc.ai/security-teams.md): Markdown mirror
- [Security (Markdown)](https://safe-doc.ai/security.md): Markdown mirror
- [Shadow Ai (Markdown)](https://safe-doc.ai/shadow-ai.md): Markdown mirror
- [Signup (Markdown)](https://safe-doc.ai/signup.md): Markdown mirror
- [Why A Layer (Markdown)](https://safe-doc.ai/why-a-layer.md): Markdown mirror