OpenAI compatible API. Attested gateway. Public status.

Providers

Provider transparency for model compute, retention, confidential compute, and encrypted routes.

Verify gateway

1 URLbase_url migration

100smodels and routes

0prompt logs by default

Gateway boundaryTLS terminates inside TrustedRouter's attested gateway.

Provider computeEach upstream provider has its own retention and enclave posture.

No prompt logsTrustedRouter stores metadata and billing records, not prompts or outputs.

API metadataMachine readable JSON remains at https://api.trustedrouter.com/v1/providers.

Provider transparency

Trusted gateway first, then explicit provider claims.

Zero data retention, confidential compute, and encrypted provider routes are shown only when explicitly tracked. Unknown stays unknown.

Provider	Trust tier	Prepaid	BYOK	ZDR	Confidential compute	Provider E2EE	Policy note
Tinfoil `tinfoil`	Confidential	yes	yes	yes	yes	yes	Tracked as a confidential inference provider with attested provider compute and no prompt/output logging claims. Policy source
Venice `venice`	Confidential	yes	yes	yes	yes	yes	Tracked as confidential — Venice documents no logging or storage of prompts/responses plus TEE-isolated, end-to-end-encrypted inference. (Caveat: requests Venice proxies to external frontier models inherit those providers' policies; TR routes Venice-native open models here.) Policy source
TrustedRouter `trustedrouter`	TR gateway	yes	yes	yes	yes	yes	TrustedRouter's attested gateway stores no prompt or output content. Provider compute policy still depends on the selected upstream route. Policy source
Anthropic `anthropic`	No logs	yes	yes	yes	not claimed	not claimed	Marked ZDR via TrustedRouter's arrangement — zero retention is NOT Anthropic's public default; it applies to contracted / approved API usage, which TrustedRouter's deployed account is configured for. Anthropic does not train on API content. (Flagged content may be retained longer for Usage-Policy enforcement; non-Messages features may differ.) Policy source
OpenAI `openai`	No logs	yes	yes	yes	not claimed	not claimed	Marked ZDR for TrustedRouter's configured OpenAI / ChatGPT API account. This is not the public OpenAI API default for every account or endpoint; ZDR depends on eligible endpoint usage and account configuration. Policy source
Gemini `gemini`	No logs	yes	yes	yes	not claimed	not claimed	Marked ZDR for TrustedRouter's Google Gemini / Vertex generative-AI routes under Google's data-governance commitments for customer prompts and outputs. Policy source
Cerebras `cerebras`	No logs	yes	yes	yes	not claimed	not claimed	Tracked as provider-ZDR. Cerebras documents ZDR-compliant ephemeral prompt caching and no persisted prompt cache data. Policy source
DeepSeek `deepseek`	No provider claim	yes	yes	no	not claimed	not claimed	Not ZDR. DeepSeek's published privacy policy says prompts/inputs may be collected and personal data may be used to train or improve machine learning models and algorithms. Policy source
Mistral `mistral`	No provider claim	yes	yes	not claimed	not claimed	not claimed	No provider-ZDR claim is tracked here. This is separate from any no-training or enterprise retention commitments Mistral may offer. Policy source
Kimi `kimi`	No provider claim	yes	yes	not claimed	not claimed	not claimed	No provider-ZDR claim is tracked here. Kimi/Moonshot policy source is linked for users who need to review API retention and processing terms. Policy source
Z.AI `zai`	No provider claim	yes	yes	not claimed	not claimed	not claimed	No provider-ZDR claim is tracked here. Z.AI/BigModel policy source is linked for users who need to review API retention and processing terms. Policy source
Together `together`	No logs	yes	yes	yes	not claimed	not claimed	Marked ZDR via TrustedRouter's arrangement — Together's ZDR is an opt-in account/privacy setting, NOT the public default, and the deployed Together account has it enabled. Together does not train on content without opt-in. Policy source
xAI Grok `grok`	No provider claim	yes	yes	not claimed	not claimed	not claimed	xAI documents no training on API requests and 30-day default audit retention, with ZDR as an enterprise feature. Policy source
Novita AI `novita`	No provider claim	yes	yes	not claimed	not claimed	not claimed	No provider-ZDR claim is tracked here. Novita's privacy policy says personal information is not used for model training; customer-content processing is governed by customer agreements. Policy source
Phala `phala`	Confidential	yes	yes	yes	yes	yes	Tracked as a confidential AI provider with provider-side attestation and encrypted prompt transport. Policy source
SiliconFlow `siliconflow`	No provider claim	yes	yes	not claimed	not claimed	not claimed	No provider-ZDR claim is tracked here. SiliconFlow's privacy policy source is linked for retention and interaction-data terms. Policy source
Parasail `parasail`	No provider claim	yes	yes	not claimed	not claimed	not claimed	Parasail documents no input logging/storage for serverless and dedicated service paths, with different handling for batch service. Policy source
Lightning AI `lightning`	No provider claim	yes	yes	not claimed	not claimed	not claimed	No provider-ZDR claim is tracked here. Lightning's general privacy and security documentation is linked for retention review. Policy source
GMI Cloud `gmi`	No provider claim	yes	yes	not claimed	not claimed	not claimed	GMI runs isolated/VPC GPU inference, but that is network isolation, NOT an attested TEE — so no confidential-compute, zero-retention, or E2EE claim is marked. Retention/training terms are unverified (the published policy page is JavaScript-only and would not render). Policy source
DeepInfra `deepinfra`	No logs	yes	yes	yes	not claimed	not claimed	Tracked as provider ZDR — DeepInfra documents memory-only handling with no storage of API content and no training on submitted API data. (Exception: requests to Google/Anthropic-backed models inherit those vendors' policies.) Policy source
Nebius Token Factory `nebius`	No logs	yes	yes	yes	not claimed	not claimed	Marked ZDR via TrustedRouter's arrangement — Nebius RETAINS inputs/outputs by default (for speculative decoding); zero retention is an opt-in control, which the deployed Nebius account has enabled. Nebius does not train on customer data. Policy source
MiniMax `minimax`	No provider claim	yes	yes	not claimed	not claimed	not claimed	No provider-ZDR claim is tracked here. MiniMax's product privacy overview is linked for users who need to review API/open-platform terms. Policy source
Xiaomi MiMo `xiaomi`	No provider claim	yes	yes	not claimed	not claimed	not claimed	No provider-ZDR claim is tracked here. Xiaomi MiMo's open-platform terms are linked for users who need to review API data handling. Policy source
Cohere `cohere`	No logs	yes	yes	yes	not claimed	not claimed	Marked ZDR — Cohere does not retain prompt/response content for TrustedRouter's configured account and does not train on customer API data. (Not a confidential-compute/TEE provider.) Policy source
Voyage AI `voyage`	No logs	yes	yes	yes	not claimed	not claimed	Marked ZDR — Voyage AI does not retain prompt content for TrustedRouter's configured account and does not train on customer API data. (Not a confidential-compute/TEE provider.) Policy source

What TR proves

The router path is verifiable.

The hosted API publishes source commit, image reference, and digest for the attested gateway. The control plane does not terminate prompt traffic.

What providers prove

Provider policy is separate.

If compute runs on Anthropic, OpenAI, Gemini, Phala, Tinfoil, or another upstream, that provider's policy controls model handling.

Routing

Choose privacy posture explicitly.

Use trustedrouter/zdr for zero-retention providers first, trustedrouter/eu for EU-focused provider selection, trustedrouter/e2e for confidential + E2EE routes, or provider routing filters for custom policy. Use trustedrouter/auto when breadth and health are the priority.