Backfrom Phala’s own policies · verified 2026-09-02 · some facts not published

Phala: data residency & retention

Headquarters
🇺🇸 United States (US)
Where inference runs
Not disclosed
No data center locations or residency statements published. Inference runs inside TEEs (trusted execution environments): 'Computation occurs within the secure TEE, and RedPill does not access memory, CPU state, or runtime data' and 'Prompts, inputs, and outputs exchanged with TEE-deployed models are end-to-end encrypted' — but no geographic disclosure.
Region pinning
Not disclosed
No region-pinning or geographic controls mentioned in privacy policy or terms.
Prompt retention
Not retained
'When using LLMs via RedPill in TEE production mode, we do not collect or store any user prompt or request data.' IP logs kept a 'maximum of 6 months for anti-spam and abuse prevention'. Claim applies to TEE production mode specifically.
Trains on API data
Not disclosed
Not explicitly addressed. ToS has generic 'We may use Content to provide, maintain, develop, and improve our Services', but in TEE mode they state prompts are not collected or stored at all.
Zero data retention
Available
Zero prompt retention is the default in TEE production mode per their privacy policy; enforced by confidential-compute architecture rather than an opt-in flag.

Compiled from Phala’s published privacy policy, terms, and documentation on 2026-09-02. “Not disclosed” means Phala publishes nothing on the point. Policies change without notice, so verify against the sources above before relying on this for compliance decisions.

Standard Compute

Plug in Standard Compute wherever you need intelligence.

One key. Every frontier model. A bill that stays put.

7-day fair refund Cancel anytime Billing by Stripe