Backfrom DeepInfra’s own policies · verified 2026-09-02
DeepInfra: data residency & retention
Headquarters
🇺🇸 United States (US)
Where inference runs
US
Inference runs in eight secure US-based datacenters DeepInfra owns and operates ('more locations rolling out globally'); specific cities are not disclosed.
Region pinning
Enterprise plans only
No region pinning or EU endpoint for the serverless API; only the enterprise DeepCluster dedicated-cluster product offers a preferred region.
Prompt retention
Not retained
API inputs and outputs are not stored to disk and exist only in memory during the request; carve-outs for discretionary debug/security sampling, image outputs, and batch API held for a short undisclosed period.
Trains on API data
No
Submitted data is not used for training models — except when using Google or Anthropic models, where the receiving company's training policy applies.
Zero data retention
Available
De-facto ZDR is the default for content (contractual no-retain clause), but DeepInfra does not brand a formal ZDR program or toggle.
Compiled from DeepInfra’s published privacy policy, terms, and documentation on 2026-09-02. “Not disclosed” means DeepInfra publishes nothing on the point. Policies change without notice, so verify against the sources above before relying on this for compliance decisions.