Claude Cowork
Pseudonymize in Claude Cowork — without exposing real data.
The Safe-Doc skill for Claude Cowork pseudonymizes your documents before Claude analyzes them. Same engine as the web UI: 55+ entities, EU self-hosted processing, zero third-party AI, zero storage.
How it works
- Install the Safe-Doc skill in Claude Cowork.
- Point to a document from Cowork — it is processed by Safe-Doc (EU servers, self-hosted models) and returns pseudonymized.
- Work in Cowork on the pseudonymized version — Claude only sees pseudonymized data.
- Restore on your machine via the API: the result is saved on your workstation, next to the original file.
Why it's safe
Claude (Cowork) never receives your real data — only the pseudonymized version. Real values only pass through two places: your workstation, and our European servers (in-memory processing, ephemeral, self-hosted models, zero third-party AI). Restoration happens locally, next to the original — no lock-in.
« Zero third-party AI » applies to our processing engine: no external AI provider sees your content during pseudonymization. You then use Cowork — and Claude only sees pseudonymized data.
User guide
Step-by-step instructions for installing and using the Safe-Doc skill in Claude Cowork.
Before you start
- Claude Cowork installed on your workstation.
- An active Safe-Doc account (14-day Solo trial or paid plan).
- The Safe-Doc skill package — request it at contact or by email at contact@safe-doc.ai. We send the installation files and setup notes with your package.
1. Install the Safe-Doc skill
Open Claude Cowork settings and add the Safe-Doc skill from the package we provide. Follow the Cowork-specific install steps included in the package (skill path, API key binding). Restart Cowork if the skill does not appear immediately.
2. Pseudonymize a document
From a Cowork session, ask Claude to run the Safe-Doc skill on a file you designate (path or document in your workspace). The file is sent to Safe-Doc, processed in memory on our European servers, and the pseudonymized version is returned to Cowork. Review the masked entities if needed before continuing.
3. Work on pseudonymized data in Cowork
Continue your session normally: summaries, rewrites, Q&A, translations. Claude only sees tokens (e.g. [PERSON_1], [ORG_2]) — not names, amounts or identifiers from the original document.
4. Restore the result locally
When you need the final document with real values back, use the restore step (via the skill or Safe-Doc API). The output is written on your workstation, next to the original file. The mapping stays under your control; nothing is archived at Safe-Doc after processing.
Troubleshooting
- Skill not listed — reinstall from the package, check the path in Cowork settings, restart the app.
- API authentication error — verify your API key in the Safe-Doc dashboard (app.safe-doc.ai); API access requires Cabinet+ or Enterprise for programmatic use.
- Need help — contact us or write to contact@safe-doc.ai.
Skill distribution: contact us to receive the installation package and Cowork setup guide.