AI Provider Trust Registry evidence verified as of 2026-08-19

Registry / Retention / ZDR

Which AI providers offer zero data retention?

As of 19 August 2026, of 16 AI model offerings tracked, 6 have clear public evidence, 9 are conditional or gated, and 1 has no public evidence.

  • Clear public evidence: Claude via Google Vertex AI, Gemini via Vertex AI, AWS Bedrock (platform), Llama via AWS Bedrock, Llama via Azure AI, DeepSeek via Fireworks AI
  • Conditional, sales-gated, or partial: OpenAI API, Anthropic API, Claude via AWS Bedrock, Mistral La Plateforme, Mistral via Azure AI, Cohere API, Cohere via AWS Bedrock, xAI API, DeepSeek API (first-party)
  • No public evidence: Azure OpenAI Service

Documented retention windows and zero-data-retention options, including whether ZDR is default, configurable, or approval-gated. The cell answers: Is retention documented, and is zero-data-retention available? Statuses below are evidence grades, not endorsements, “no public evidence” means we could not verify it from public sources, not that the answer is no.

OpenAI API first-party API
Yes, sales-gated confidence: high · verified 2026-08-17

OpenAI documents a 30‑day default retention and states Zero Data Retention is only available after sales‑gated approval, matching yes_sales_gated.

source · full cell

Azure OpenAI Service OpenAI model, served by Microsoft Azure
No public evidence confidence: low · verified 2026-08-15

Official Microsoft documentation of retention and zero-data-retention is behind an authentication gate, so no publicly accessible evidence exists.

source · full cell

Anthropic API first-party API
Yes, sales-gated confidence: high · verified 2026-08-19

Anthropic’s official docs publicly describe data retention and state ZDR is available only after contacting the sales team, matching yes_sales_gated.

source · full cell

Claude via AWS Bedrock Anthropic model, served by AWS Bedrock
Partial confidence: high · verified 2026-08-14

Retention is publicly documented and self-serve configurable, but ZDR access for Claude models is gated behind AWS/Anthropic account representative approval, fitting partial per the rubric.

source · full cell

Claude via Google Vertex AI Anthropic model, served by Google Cloud Vertex AI
Yes, public confidence: high · verified 2026-08-16

The provider’s own data‑governance page documents retention policy and describes how to achieve zero data retention for all managed models, which includes partner models like Claude via Vertex AI.

source · full cell

Gemini via Vertex AI Google model, served by Google Cloud Vertex AI
Yes, public confidence: high · verified 2026-08-16

The Google Cloud page publicly documents data retention policies and provides specific actions for achieving zero data retention for Gemini via Vertex AI.

source · full cell

AWS Bedrock (platform) platform row
Yes, public confidence: high · verified 2026-08-19

The AWS Bedrock documentation publicly describes a zero‑data‑retention mode, confirming the retention policy is documented and the feature is available.

source · full cell

Mistral La Plateforme first-party API
Yes, sales-gated confidence: high · verified 2026-08-19

The page publicly documents retention control and states ZDR is only available via a gated pay‑as‑you‑go plan and request, matching yes_sales_gated.

source · full cell

Mistral via Azure AI Mistral AI model, served by Microsoft Azure
Yes, sales-gated confidence: high · verified 2026-08-15

The recorded source was about Azure OpenAI (OpenAI's models), not Mistral via Azure AI. The correct source is Microsoft's "Data, privacy, and security for Models sold by Azure" page, which covers Mistral AI models sold through Azure. It publicly documents retention for abuse monitoring and states th

source · full cell

Cohere API Cohere model, served by Cohere (first-party)
Yes, sales-gated confidence: high · verified 2026-08-19

Cohere publicly documents a 30‑day retention policy and states zero‑data‑retention is available only after approval, matching a yes_sales_gated classification.

source · full cell

Cohere via AWS Bedrock Cohere model, served by AWS Bedrock
Partial confidence: high · verified 2026-08-17

The AWS Bedrock data‑retention page publicly documents retention policies and states ZDR is only available after contacting an account manager, so the claim is only partially satisfied.

source · full cell

Llama via AWS Bedrock Meta model, served by AWS Bedrock
Yes, public confidence: high · verified 2026-08-14

AWS Bedrock documentation publicly describes data‑retention settings and notes that zero‑data‑retention mode exists, though eligibility is evaluated per account and model.

source · full cell

Llama via Azure AI Meta model, served by Microsoft Azure (Azure AI Foundry / Models-as-a-Service)
Yes, public confidence: high · verified 2026-08-17

The cited Azure OpenAI "Models sold by Azure" page does not cover Llama (a third-party catalog model); the authoritative Microsoft page for Llama-on-Azure is the model-catalog data-privacy page (learn.microsoft.com/en-us/azure/foundry-classic/how-to/concept-data-privacy), which documents statelessne

source · full cell

xAI API xAI model, served by xAI (first-party)
Partial confidence: high · verified 2026-08-14

Retention is publicly documented, while zero‑data‑retention is only available via a self‑serve toggle that may not be enabled for all customers, making the overall answer partially satisfied.

source · full cell

DeepSeek API (first-party) first-party API
Partial confidence: high · verified 2026-08-16

The privacy policy documents data retention, but no public page from DeepSeek describes a zero‑data‑retention offering, so only part of the claim is satisfied.

source · full cell

DeepSeek via Fireworks AI DeepSeek model, served by Fireworks AI
Yes, public confidence: high · verified 2026-08-14

Fireworks publicly documents zero-data-retention for open models and DeepSeek is classified as an open model, confirming retention is documented and zero-data-retention is available.

source · full cell