How do we get zero data retention for the Claude API?

Going direct, Anthropic enables zero data retention per organization through its account team after you contact sales, and each new organization needs it enabled separately. Through Standard Compute, an admin turns on the requirement for the company, a team or a key, and every request on those keys goes only to endpoints that commit to keeping no copy of your data.

Does OpenAI keep our API data?

OpenAI's documentation says abuse monitoring logs are generated for all API usage and kept for up to 30 days by default, and that zero data retention is subject to prior approval by OpenAI.

Can a team or a developer switch zero data retention off?

No. Once the company requires zero data retention, teams and keys inherit the requirement and can never switch it off underneath you.

Does Standard Compute store our prompts?

No. We process your prompts in memory to serve each request and never write them to storage or logs. Request records hold metadata only (model, provider, tokens, cost and status), never your prompts or outputs.

How do we show this in a vendor review?

Sign the Data Processing Agreement in the dashboard: it commits us to no training on your data, no retention of your content, and breach notification within 72 hours. The admin audit log records every change to your rules with before and after values, and the request export shows the model, the provider that served it and the region rule in force for every request, kept for 12 months.

Does zero data retention work with open models?

Yes. The requirement applies to every model on the key, including open models such as GLM, DeepSeek and MiniMax next to Claude, GPT and Gemini.

Zero data retention for Claude, GPT and open models at company level

Standard Compute makes zero data retention one setting across Claude, GPT, Gemini, Mistral, Grok, DeepSeek, Qwen, GLM, MiniMax, Kimi and MiMo. Turn it on for your company, a team or a key, and every request is routed only to inference endpoints that commit to keeping no copy of your data. Teams and keys inherit the requirement and can never switch it off underneath you.

Going direct, each vendor decides. Anthropic enables zero data retention per organization through its account team, OpenAI requires prior approval, and Mistral reviews each request at its discretion. A company that uses all three makes three requests and keeps three sets of exceptions.

By Standard Compute · Last updated 9 November 2026

How you get zero data retention, vendor by vendor

What each vendor's own documentation says, next to how Standard Compute does it.

How you get zero data retention, vendor by vendor
Route	How it is switched on	What the vendor says it covers
Anthropic Claude API	How it is switched on: Per organization, enabled by your Anthropic account team after you contact sales	What the vendor says it covers: Messages and Token Counting for eligible features; Claude Team and Claude Enterprise interfaces are not eligible
OpenAI API	How it is switched on: Prior approval by OpenAI and acceptance of additional requirements	What the vendor says it covers: Abuse monitoring logs are kept for up to 30 days by default; chat completions and responses are eligible with limitations
Mistral API	How it is switched on: A request through Mistral's Help Center, approved or denied at Mistral's discretion	What the vendor says it covers: Pay-as-you-go, stateless API calls only; agents, conversations and libraries are excluded
Standard Compute	How it is switched on: A setting your admin turns on for the company, a team or a key	What the vendor says it covers: Every request is routed only to inference endpoints that commit to keeping no copy of your data

Sources: each vendor's own documentation, listed at the end of this page and checked on 9 November 2026.

Create a company workspaceSee pricing

How do we get zero data retention for Claude and OpenAI at company level?

Create a company workspace, turn on the zero data retention requirement, and give every person and every app its own key. The requirement then holds on every request those keys send, whichever model or provider serves it.

Settings combine down the hierarchy. A team or a key can add the requirement when the company has not, and once the company requires it, no team or key can loosen it. Region rules and the zero data retention requirement are free.

How do we get zero data retention for Claude and OpenAI at company level?
Level	Effect of the requirement
Company	Effect of the requirement: Applies to every team and every key
Team	Effect of the requirement: Inherits the company setting, or adds the requirement for that team only
Key	Effect of the requirement: Inherits from the team and company, or adds the requirement for one app or person

An AI API that keeps no copy of your prompts

We process your prompts in memory to serve each request and never write them to storage or logs.

Your account and usage records are stored in Frankfurt, Germany, and our EU gateway runs in Finland. Request records hold metadata only (model, provider, tokens, cost and status), never your prompts or outputs.

Standard Compute never trains on your prompts or outputs. With a region rule or zero retention turned on, every request goes only to providers that commit not to train on it.

Sign our Data Processing Agreement in the dashboard in minutes. It commits us to no training on your data, no retention of your content, and breach notification within 72 hours. We publish every sub-processor with its purpose, data and location, and give you 14 days' notice before adding a new one.

Zero data retention for Claude Code and Claude Desktop

Anthropic lists the Claude Team and Claude Enterprise interfaces as not eligible for its API zero data retention arrangement, and offers a separate arrangement for Claude Code on Claude Enterprise to qualified accounts.

With Standard Compute, Claude Code and Claude Desktop use a key from your company workspace, so the zero data retention requirement on that key applies to every request they send. Point Claude Code and Claude Desktop at Standard Compute and your people stop hitting 5-hour and weekly caps: agents run until the monthly budget is used.

export ANTHROPIC_BASE_URL=https://eu.stdcmpt.com
export ANTHROPIC_AUTH_TOKEN=KEY_FROM_YOUR_COMPANY_WORKSPACE
claude

Claude Code on a company key

Zero data retention and processing in Europe together

Many companies need both. Set the zero data retention requirement and a Europe rule on the same company, team or key.

Set a Europe rule for your company, a team or a single key, and every request, including retries and fallbacks, is served only by approved European inference endpoints. If no European endpoint can serve it, the request is refused before your prompt is sent.

Our EU gateway terminates TLS and runs in Finland, and your account data is stored in Frankfurt.

Questions from IT, security and finance

How do we get zero data retention for the Claude API?: Going direct, Anthropic enables zero data retention per organization through its account team after you contact sales, and each new organization needs it enabled separately. Through Standard Compute, an admin turns on the requirement for the company, a team or a key, and every request on those keys goes only to endpoints that commit to keeping no copy of your data.
Does OpenAI keep our API data?: OpenAI's documentation says abuse monitoring logs are generated for all API usage and kept for up to 30 days by default, and that zero data retention is subject to prior approval by OpenAI.
Can a team or a developer switch zero data retention off?: No. Once the company requires zero data retention, teams and keys inherit the requirement and can never switch it off underneath you.
Does Standard Compute store our prompts?: No. We process your prompts in memory to serve each request and never write them to storage or logs. Request records hold metadata only (model, provider, tokens, cost and status), never your prompts or outputs.
How do we show this in a vendor review?: Sign the Data Processing Agreement in the dashboard: it commits us to no training on your data, no retention of your content, and breach notification within 72 hours. The admin audit log records every change to your rules with before and after values, and the request export shows the model, the provider that served it and the region rule in force for every request, kept for 12 months.
Does zero data retention work with open models?: Yes. The requirement applies to every model on the key, including open models such as GLM, DeepSeek and MiniMax next to Claude, GPT and Gemini.