Short answer: it’s not automatic. You must (a) use the EU “isolated environment” for data residency and (b) explicitly enable Zero Retention Mode.
EU data residency (Enterprise only)
Request access to the EU isolated environment, then use the EU URLs:
WebSocket base: wss://api.eu.residency.elevenlabs.io
Use a separate EU account/API key and point your SDK/base_url to the EU API. Zero Retention is not auto‑enabled in isolated environments. (elevenlabs.io)
Note: By default, storage is in the selected region; some processing may occur outside the EU. To restrict processing to the EU, use the EU isolated environment together with Zero Retention Mode via the API. (elevenlabs.io)
Zero data retention for Scribe (API)
Zero Retention Mode is an Enterprise feature. It prevents storage/logging of STT inputs/outputs when enabled. (elevenlabs.io)
For Scribe’s REST transcription endpoint, pass enable_logging=false on the request. This disables logs and transcript storage for that request. Example: POST /v1/speech-to-text/convert?enable_logging=false. (elevenlabs.io)
General ZRM behavior and enablement are documented here (including per‑request enable_logging=false and Enterprise availability). (elevenlabs.io)
Scribe v2 support
Scribe v2 Realtime is “enterprise‑ready” and supports EU (and India) data residency options and Zero Retention Mode; use the EU isolated environment above and ZRM as described. (elevenlabs.io)
Bottom line: Use the EU isolated environment endpoints for residency, and set enable_logging=false for Scribe API calls to achieve zero data retention. Neither is enabled by default. (elevenlabs.io)
echo kite
calm dove