Skip to content

FAQ

Nowhere outside your infrastructure. The engine runs on your side, the token-to-real-value mapping is encrypted at rest on your server, and the AI provider only receives pseudonymized text it cannot re-identify. Full detail: Security and technical measures.

No. The software runs on your infrastructure, the vendor has no access to the data and no telemetry carries it. TokenVeil adds no processor to your records. The GDPR dossier details the roles, article by article.

Pseudonymization, in the sense of GDPR Article 4(5): the replacement is reversible on your side (that is what restores the response), irreversible on the AI provider’s side (it does not hold the mapping). We say “anonymization” in the interface for simplicity; the compliance dossier uses the rigorous term.

What is the difference between Community and Enterprise?

Section titled “What is the difference between Community and Enterprise?”

The detection engine, nothing else. Community: regex, deterministic data, free. Enterprise: ML (Presidio + spaCy NER), free-text names, documents, LDAP, multi-tenancy. Full comparison: Community vs Enterprise.

Yes. The method is public and reproducible (annotated test set plus random-seed fuzzing, independent verifier, continuous-integration gate). See the benchmark page; the commands to re-run it ship with every instance.

Claude (direct, Vertex AI or Bedrock), Gemini, OpenAI, Mistral, GitHub Models, Azure OpenAI. Model lists update automatically from each provider.

Yes: the Community edition is free, with no seat lockout, and the code is public. Elastic License 2.0: you can read, audit, self-host and modify; not resell it as a managed service.

Request a demo or write to contact@tokenveil.eu. Commercial license, on-prem deployment, support.