Skip to main content
Watch: I Built an AI App Without Leaving Databricks

Data retention - verified

Does Mistral store your API prompts?

Default retention: 30 rolling days, abuse monitoring. Trains on your API data: Not stated (opt-out toggle exists; the default for paid API use is not stated). Zero data retention: by approval (on request; Mistral may approve or deny). Regional processing: US and EU. Mistral's own pages disagree on who can get ZDR; both versions are quoted below.

Read from legal.mistral.ai, docs.mistral.ai, help.mistral.ai. Compare every AI API's data retention

Trains on data
Not stated
ZDR
By approval
Regions
US and EU
Verified
2026-10-18

The docs disagree

Mistral's own pages say different things here. Both are quoted; neither is picked. Ask Mistral which applies to your account.

The docs disagree

Who can get ZDR

Mistral's docs say ZDR is available on paid plans. Its help center says ZDR is available only with pay-as-you-go. Both say stateless API calls only.

“ZDR is available on paid plans for supported stateless API calls.”
Mistral docs - docs.mistral.ai
“Zero Data Retention (ZDR) is available only with pay-as-you-go and only for stateless API calls.”
Mistral help center - help.mistral.ai

How long are prompts kept by default?

Default retention

30 rolling days, abuse monitoring

API inputs and outputs kept for the time needed to generate the output, then 30 rolling days for abuse monitoring, unless ZDR is active. Some APIs (Agents, Fine-Tuning) keep data longer.

“Except for specific APIs, we keep your Input and Output for the period necessary to generate the Output and then for thirty (30) rolling days to monitor abuse (unless zero data retention is activated).”
legal.mistral.ai
Duration
Not stated
Purpose
Not stated

Does it train on your API data?

Not stated

Opt-out toggle exists; the default for paid API use is not stated

Commercial terms: no training on Customer Data or Outputs except where the customer opted in, has not opted out of a product set to opt-in by default, gave feedback, an Order Form says so, or uses Labs/Preview models. Help center: Free mode data may be used for training (opt-out available); pay-as-you-go customers can opt out at any time. The default state of the API toggle for pay-as-you-go is not stated explicitly.

“4.2. Training. Mistral AI will not use Customer Data or Outputs to train its artificial intelligence models except (a) when you (i) opted-in to training on a Mistral AI Product set to opt-out by default or (ii) have not opted-out of training on a Mistral AI Product set to opt-in by default, (b) when Customer or an End User provides Feedback to Mistral AI, (c) as otherwise may be provided in an Order Form or (d) when Customer uses Labs or Preview Models.”
legal.mistral.ai
“As stated during subscription, we may use your data (input and output) to train our artificial intelligence models.”
help.mistral.ai
“Under the `Anonymous improvement data` section, disable the toggle to prevent API calls and related data from being used to improve Mistral’s services.”
help.mistral.ai

Zero data retention (ZDR)

Availability

By approval: the provider must approve your request

On request; Mistral may approve or deny

Who can get it

Docs: paid plans, supported stateless API calls. Help center: only with pay-as-you-go. Granted on request with a legitimate reason; Mistral may approve or deny. Not available for Labs models.

Not covered

Agents, batch processing files, Conversations, Libraries, /v1/files, Vibe Work, Chat; Labs models.

From the docs
“ZDR is available on paid plans for supported stateless API calls.”
docs.mistral.ai
“When ZDR is enabled, Mistral does not store or log inputs and outputs for supported API requests longer than required to generate the output.”
docs.mistral.ai
“ZDR applies to supported endpoints across models, except Labs models.”
docs.mistral.ai
“Zero Data Retention (ZDR) is available only with pay-as-you-go and only for stateless API calls.”
help.mistral.ai
“We will review your request and, at our discretion, approve or deny it.”
help.mistral.ai

Features kept longer

Stateful features hold your content until a job finishes, a window expires or you delete it, often regardless of the default above.

Mistral features with their own retention
FeatureRetentionThe docs say
Agents APIUntil account termination“If you use our Agents API, we keep your Input and Output until you terminate your account.”legal.mistral.ai
Fine-Tuning APIUntil deleted from Mistral AI Studio or account termination“If you use our Fine-Tuning API, we keep your fine-tuning data until you delete it from Mistral AI Studio or until you terminate your account.”legal.mistral.ai
Files API, batch files, Conversations, Libraries (stateful, outside ZDR)Not stated“For example, the Files API stores uploaded files and is outside the scope of ZDR.”docs.mistral.ai
Labs or Preview ModelsNot stated“By using Labs or Preview Models, you acknowledge that (i) Mistral AI may use Customer Data and Outputs generated from Labs or Preview Models to train its artificial intelligence models and (ii) the opt-out preferences you selected for other Mistral AI Products (including through zero data retention)does not apply to Labs or Preview Models.”legal.mistral.ai

Where is your data stored and processed?

Regional processing

US and EU

Regional endpoints at 1.1x list price

Stated location

No commitment on the global endpoint; EU hosting by default per the help center

Recorded as: Global endpoint: no committed location. Help center: hosted in the European Union by default.

How to choose
Not stated
What it covers

Regional endpoints: processing (inference). Help center: hosting (storage) EU by default.

Summary

Optional regional inference endpoints for the EU and the US (10% upcharge) govern where inference processing happens. The global endpoint has no committed inference location. Regional inference does not provide regional storage for all control-plane data. Separately, the help center says data is hosted in the EU by default and in the US when the US endpoint is used.

From the docs
“It supports processing inference requests by systems located within a chosen geography: currently the European Union or the United States.”
docs.mistral.ai
“Regional inference routes that model execution to infrastructure located within a selected geography. The request still uses the Mistral API, but the regional endpoint determines where inference processing happens.”
docs.mistral.ai
“Regional inference does not provide regional storage for all control-plane data. For example, account configuration, API keys, billing, access management, usage analytics, and other operational metadata may still be handled by Mistral systems outside the selected inference geography.”
docs.mistral.ai
“Stateful features, including Agents, Batch, and the Files API, are not available on regional endpoints.”
docs.mistral.ai
“By default, your data is hosted in the European Union.”
help.mistral.ai
“You may, however, explicitely use our US API endpoint and in such a case your data is hosted in the United States.”
help.mistral.ai
“Depending on the feature you use, your data can be temporarily transferred outside of the European Union, to the locations listed in the “Subprocessors” tab of our Trust Center.”
help.mistral.ai

Regional endpoints

  • Global
    api.mistral.ai
    “Use the global endpoint when you do not need a specific inference location. Mistral does not commit to a specific inference location for requests sent to this endpoint.”
    docs.mistral.ai
  • EU
    api.eu.mistral.ai
    “Multiple data centers in EU and EFTA countries”
    docs.mistral.ai
  • US
    api.us.mistral.ai
    “Multiple data centers in the United States”
    docs.mistral.ai

Certifications

Certifications on the pages checked: Not stated

Research notes

  • Regional inference billed at 1.1x list price. DPA: personal data no longer accessible 30 days after termination of access. Feedback (thumbs up/down) authorizes training use on any plan. Help center wording on ZDR eligibility (pay-as-you-go only) is narrower than the docs (paid plans).

FAQ

Does Mistral store your API prompts?

30 rolling days, abuse monitoring. API inputs and outputs kept for the time needed to generate the output, then 30 rolling days for abuse monitoring, unless ZDR is active. Some APIs (Agents, Fine-Tuning) keep data longer. The provider says: "Except for specific APIs, we keep your Input and Output for the period necessary to generate the Output and then for thirty (30) rolling days to monitor abuse (unless zero data retention is activated)."

Does Mistral train on your API data?

Not stated (opt-out toggle exists; the default for paid API use is not stated). The provider says: "4.2. Training. Mistral AI will not use Customer Data or Outputs to train its artificial intelligence models except (a) when you (i) opted-in to training on a Mistral AI Product set to opt-out by default or (ii) have not opted-out of training on a Mistral AI Product set to opt-in by default, (b) when Customer or an End User provides Feedback to Mistral AI, (c) as otherwise may be provided in an Order Form or (d) when Customer uses Labs or Preview Models."

Does Mistral offer zero data retention?

By approval (on request; Mistral may approve or deny). Docs: paid plans, supported stateless API calls. Help center: only with pay-as-you-go. Granted on request with a legitimate reason; Mistral may approve or deny. Not available for Labs models.

Where does Mistral process your data?

Regional processing options: US and EU (regional endpoints at 1.1x list price). Stated processing location: No commitment on the global endpoint; EU hosting by default per the help center.

Sources

Last verified from Mistral's own pages below. Quotes are verbatim.

This page summarizes what each provider's own pages say. It is not legal advice; your contract, data processing addendum and the provider's current terms govern what happens to your data.

Newsletter

Know when a provider changes its data terms

One email per week. New model releases, pricing and policy changes, and what they mean for what you build.