Key Facts
What must Customer disclose where required by applicable law?
Mistral AI requires Customer to disclose AI-generated or partially AI-generated audio content where required by applicable law.
What must Customer comply with by using audio Mistral AI Products?
Mistral AI requires Customer, by using audio Mistral AI Products, to comply with all applicable laws and Mistral AI's Usage Policy.
What must Customer do upon Mistral AI's request?
Mistral AI requires Customer, upon Mistral AI's request, to reasonably cooperate with Mistral AI to implement any security or safety measure to mitigate or remediate any systemic risks in accordance with Applicable AI Laws.
Why must Customer cooperate with Mistral AI?
Mistral AI requires Customer, upon Mistral AI's request, to reasonably cooperate with Mistral AI to implement any security or safety measure to mitigate or remediate any systemic risks in accordance with Applicable AI Laws.
What must Customer notify Mistral AI of?
Mistral AI requires Customer to notify Mistral AI of any Serious Incident within forty-eight hours of becoming aware of it.
What is prohibited regarding End User Accounts?
Mistral AI prohibits End User Accounts from being shared, sold, licensed, or otherwise made available outside of Customer's entity or organization.
What use is prohibited for Media Customers using the verified news feature?
Mistral AI prohibits Media Customers from using the verified news feature in Mistral AI Products for editorial use.
What does Mistral AI exclude its liability for?
Mistral AI excludes its liability for any damages, liabilities, losses, charges, or expenses arising out of or in connection with issues related to Customer Exportable Data and Assets integrity, loss, compatibility issues, or any other issue occurring during or as a result of the Switch.
What may Mistral AI use Usage Data for?
Mistral AI may use Usage Data for its business purposes, including to conduct research, improve the Mistral AI Products, create new Mistral AI Products, or enhance product performance, functionality, and user experience.
What is Customer prohibited from including in its Inputs or purposefully generating in Outputs?
Mistral AI prohibits Customer from including in its Inputs, or purposefully generating Outputs, content that violates applicable law or the Usage Policy.
Summary
These terms set out the rules Customers must follow when using Mistral AI's products, including audio tools and web-search features. You are responsible for reporting serious problems within 48 hours, keeping End User Accounts within your own organization, and not using AI-generated audio or web-search content in prohibited ways. If you move data away from Mistral AI (a 'Switch'), Mistral AI is not responsible for any data loss or problems that occur, and you may be charged a fixed fee plus an hourly rate for that process.
Analysis
This document establishes a set of product-specific obligations, restrictions, and risk allocations governing Customer use of Mistral AI's audio and other AI products. Customers bear a dual compliance burden: adherence to applicable law and Mistral AI's Usage Policy across all product interactions, including a forty-eight-hour Serious Incident notification obligation and a demand-triggered duty to cooperate on AI-law-mandated security and safety measures. The document restricts how Customers handle End User Accounts, Inputs, Outputs, Third-Party Content from web-search, and Third-Party Service connections, and it limits Mistral AI's liability entirely for data integrity, loss, and compatibility issues arising from the Switch process. A time-limited fee schedule applies to Switching Requests made before 12 January 2027, combining a fixed EUR 1,000 charge with an uncapped hourly rate of EUR 180, both exclusive of taxes.
What this means for you
As a user of Mistral AI products, you must notify Mistral AI of any Serious Incident within forty-eight hours of first becoming aware of it, and you must not share, sell, or license your End User Account to anyone outside your organization. When you enable connections to Third-Party Services, you grant Mistral AI permission to access those services and perform a broad range of actions — including modifying, deleting, and invoking operations — on your behalf, so you should enable only those connections whose access scope you accept. Where applicable law requires it, you must disclose when audio content you publish was fully or partially AI-generated.
2 important changes detected
3 versions captured · Last updated: August 2026
What changed
Mistral AI updated its Additional Product Terms effective August 5, 2026, making operational clarifications to its third-party integration and product governance frameworks. The company renamed 'Third-Party Services' to 'Third-Party Connectors' throughout, replaced 'Inputs' with 'Customer Data' in its acceptable use policy, and introduced new definitions and liability disclaimers for 'Third-Party Products' (third-party AI models accessed through Mistral's platform). Under the revised terms, Mistral AI explicitly states it has no obligation to fulfill AI regulatory responsibilities for third-party products, and disclaims liability for them to the maximum extent permitted by law.
Why this matters
The updated terms clarify Mistral AI's governance of third-party AI models and integrations accessed through its platform. Mistral AI now explicitly states that third-party products are provided as-is without warranties, that it has no obligation to indemnify customers against third-party intellectual property claims, and that it bears no regulatory responsibility for third-party products under applicable AI laws. Output generated using third-party products remains assigned to customers. The company also renamed 'Third-Party Services' to 'Third-Party Connectors' and changed 'Inputs' to 'Customer Data' in its acceptable use policies for consistency.
View full change record →
What changed
Mistral AI updated its Additional Product Terms on May 29, 2026, shifting from 'Customer' to 'you' language throughout the Third-Party Service integration section. More substantively, the terms now authorize Mistral AI to execute a broader set of actions on data sent to third-party services, expanding from simple access and retrieval to include execute, transmit, modify, delete, invoke, and otherwise act upon necessary data. The updated language also places explicit responsibility on users for actions they request through the product, including compliance with third-party terms.
Why this matters
The updated terms authorize Mistral AI to perform a broader range of actions on data sent to third-party services, moving beyond simple access and retrieval to include execute, transmit, modify, delete, invoke, and otherwise act upon necessary data. Users now explicitly bear responsibility for the actions they request through Mistral AI Products and for compliance with third-party terms. If you use third-party service integrations, you should review what specific actions you are requesting and confirm you understand what data handling Mistral AI is authorized to perform on your behalf.
View full change record →