Enterprise Features

Updated November 2, 2026 — Le Chat is now Vibe. As of May 28, 2026, Le Chat has been rebranded as Vibe (still at chat.mistral.ai): a single agent for work and code that carries over your conversations, settings, and subscriptions. The legacy Le Chat experience remains available in Chat mode. In Work mode, agents are now skills (invoked with @), memories form the knowledge base, and deep research is the Deep Research skill. Plans are now called Mistral Free, Pro, Team, and Enterprise. This lesson uses the name Vibe; note that some menus described here may have changed names or locations.

The Real Gap Between Enterprise and Consumer Versions

Mistral Enterprise isn’t just a paid version of Vibe. It’s a platform built around what an organization needs when equipping hundreds of employees: centralized authentication, granular access administration, complete usage traceability, and contractual guarantees for data protection. The difference barely shows in the chat interface itself; it shows in everything surrounding it—precisely the scope your security leadership will examine.

Identities and Authentication

Mistral Enterprise connects natively to your organization’s existing identity provider using SSO (Single Sign-On) and SAML 2.0 protocols. Your employees log in with their usual credentials, whether from Azure AD, Okta, Google Workspace, or any other compatible IdP. The primary benefit is automatic provisioning: accounts are created and deactivated in sync with your directory, and multi-factor authentication is inherited from your IdP with no extra configuration. When an employee leaves the company, deactivating them in the directory immediately cuts off their assistant access, whereas an individual subscription might remain open for months. Audits become simpler because there’s only one access path to document.

Administration then relies on a four-level hierarchical role system. The global administrator holds the organization’s configuration, SSO management, and security policies; this role goes to only a handful of people in your IT department. The team administrator manages members and access for their specific department only, letting a marketing manager add an intern without opening a ticket. The standard user accesses features allowed by organizational policies, while the restricted user is limited to certain models or features—a status useful for external contractors.

Each role carries policies that determine accessible models, enabled features—web search, code generation, file uploads—and usage limits. This enables highly differentiated configurations within a single organization: a finance department works on the most powerful models with web search disabled, while a development team has code execution enabled under a monthly quota.

Zero Retention That Unlocks Entire Use Cases

The most structural commitment of the Enterprise plan is zero retention. No conversations are stored after the session ends, no data feeds model training or improvement, and no history remains on the server: when a conversation closes, data is permanently deleted from Mistral’s infrastructure.

This guarantee isn’t just boilerplate; it changes the list of topics your teams are allowed to discuss with the assistant. Legal teams approve usage without fearing unauthorized retention, often shortening project qualification by several weeks. Financial, medical, or strategic data becomes eligible for processing—a margin table before earnings release that no one would paste into a consumer tool. GDPR compliance becomes lighter since there’s no subsequent processing to describe in your register.

The Admin Console

The admin console serves as the deployment’s control center. Its dashboard displays real-time daily, weekly, and monthly active users, query volume by team and model, usage breakdown by type—conversation, search, document analysis, code—plus adoption trends and retention rates. These are the numbers that will tell you, six months after launch, whether the deployment has taken hold or stalled in two pilot departments.

Every interaction also feeds an audit log that records the user identity provided by SSO, a precise timestamp, the action type performed—conversation, upload, search—and the model invoked. These logs export in a standard format for ingestion into your SIEM, such as Splunk or Elastic, and cover traceability requirements for finance, healthcare, or defense. From the same console, you apply policies at the organization level: restrict certain models by team, enable content filtering guardrails adapted to your sector, set quotas per user or team, and limit access to specific IP ranges to block any connection outside the network.

Advanced Capabilities

Beyond standard conversation, four use cases appear in almost every deployment. Document analysis lets you upload PDFs, Word files, Excel spreadsheets, or images and query the assistant about their content; the extraction engine handles tables, charts, and complex layouts, which becomes critical when submitting an annual report rather than a two-page memo. Code generation comes with an integrated editor featuring syntax highlighting, execution in a secure sandbox, and support for more than 80 programming languages.

Canvas opens a dedicated space for long documents, where writing, revising, rephrasing, and translating happen iteratively: the text evolves through exchanges instead of being regenerated with every message. Administrators can also create custom agents by combining predefined system instructions, reference documents, and specific tools, then distribute them across the organization—legal agent, HR agent, technical agent. This is often where adoption happens: an HR agent that correctly answers leave questions will do more for your deployment than any internal presentation.

Key Takeaways

• SSO/SAML 2.0 ensures centralized authentication and automatic provisioning
• Zero retention guarantees no data is kept after the session ends
• The admin console offers usage metrics, exportable audit logs, and granular policies
• Advanced features (documents, code, Canvas, agents) are available in the Enterprise plan
• The role system enables fine-grained access control by team and user