Tabnine

Data-handling profile

Trains on input

No — does not train on customer inputs

Retention

No-train-no-retain policy across all plans and deployment types: code sent for inference is deleted immediately after the response is returned and never stored server-side. Completion/chat models are trained only on permissively-licensed open-source code, not customer code. Enterprise offers private, self-contained (VPC/on-prem/air-gapped) deployment.

Data region

Global

Certifications

none listed

DPA available

Yes

Breach history

none known

EU AI Act

Zero-retention-by-design and air-gapped enterprise deployment options make this one of the strongest data-governance postures in the coding-assistant category for EU AI Act and IP-sensitive use cases.

Sources

• https://www.tabnine.com/code-privacy/
• https://docs.tabnine.com/main/welcome/readme/privacy

Every fact above is drawn from the vendor's public documentation, reviewed by a human, and version-stamped. Monitored daily for changes.

Change history

No material changes recorded since monitoring began.