Mistral Expands Studio With Agents and Apps

Mistral says Mistral Studio, its enterprise AI platform, is expanding with managed agents and apps plus new workflow, enterprise context, observability and evaluation capabilities. The company says teams can connect models such as Mistral Large 4, also called Le Chonk, to business data and systems, then deploy in the cloud, a private VPC, on-premises or across environments. Mistral says customers retain control over deployment, data access and the systems’ accumulated knowledge.

Studio combines agents that investigate and act, workflows that coordinate steps and approvals, and apps employees use. Developers can connect enterprise data through connectors and MCP servers. Through a new Mistral MCP server, they can use Mistral Vibe or another coding agent, test locally and deploy to Studio with one command; Studio provides hosting and execution.

Managed agents can perform multi-step tasks using team procedures, code-running sandboxes and subagents that handle delegated tasks. Developers can add hooks (custom logic at points in a program’s execution) and set limits on agents’ tools, data and actions. Mistral says Studio uses infrastructure supporting millions of Mistral Vibe sessions each week. Its engineering team built an incident-investigation agent that follows a runbook, requests approval before sensitive actions and posts findings in Slack; Mistral says it saves engineers hundreds of hours each month.

Workflows specify which steps follow fixed rules, where an agent reasons and when a person decides. Interrupted runs can resume, and developers can generate, test and deploy workflows from process descriptions in Mistral Vibe, with code they can inspect and edit. The release adds enterprise-system event triggers and a visual graph of workflow logic. Mistral’s product operations team uses a workflow to identify bugs and feature requests in internal ideas and customer feedback, then route them to the appropriate teams; the company says this saves about 35 hours a week.

Studio-hosted apps include authentication and access controls, and teams can build with reusable, pre-approved components to reduce new code requiring security review. Apps can use agents and workflows, search enterprise documents and draw on shared Studio context. Mistral’s Applied AI team built a staffing app that recommends people for projects and explains how their experience fits; staffing leads review a shortlist with supporting evidence.

Mistral says Studio deployments can run in a customer’s cloud, VPC or on-premises environment, on Mistral-hosted infrastructure, or across a combination; sensitive data and related processes can remain within the customer’s perimeter. OpenTelemetry traces record model calls, tool use and approvals, while evaluations let teams test prompt, model or tool changes against their own datasets before release. Mistral says customers use systems built with its Solutions team for tasks from penetration testing to KYC onboarding, keeping sensitive data and specialist knowledge under their control.

Mistral says Mistral Studio, its enterprise AI platform, is expanding with managed agents and apps plus new workflow, enterprise context, observability and evaluation capabilities. The company says teams can connect models such as Mistral Large 4, also called Le Chonk, to business data and systems, then deploy in the cloud, a private VPC, on-premises or across environments. Mistral says customers retain control over deployment, data access and the systems’ accumulated knowledge.

Studio combines agents that investigate and act, workflows that coordinate steps and approvals, and apps employees use. Developers can connect enterprise data through connectors and MCP servers. Through a new Mistral MCP server, they can use Mistral Vibe or another coding agent, test locally and deploy to Studio with one command; Studio provides hosting and execution.

Managed agents can perform multi-step tasks using team procedures, code-running sandboxes and subagents that handle delegated tasks. Developers can add hooks (custom logic at points in a program’s execution) and set limits on agents’ tools, data and actions. Mistral says Studio uses infrastructure supporting millions of Mistral Vibe sessions each week. Its engineering team built an incident-investigation agent that follows a runbook, requests approval before sensitive actions and posts findings in Slack; Mistral says it saves engineers hundreds of hours each month.

Workflows specify which steps follow fixed rules, where an agent reasons and when a person decides. Interrupted runs can resume, and developers can generate, test and deploy workflows from process descriptions in Mistral Vibe, with code they can inspect and edit. The release adds enterprise-system event triggers and a visual graph of workflow logic. Mistral’s product operations team uses a workflow to identify bugs and feature requests in internal ideas and customer feedback, then route them to the appropriate teams; the company says this saves about 35 hours a week.

Studio-hosted apps include authentication and access controls, and teams can build with reusable, pre-approved components to reduce new code requiring security review. Apps can use agents and workflows, search enterprise documents and draw on shared Studio context. Mistral’s Applied AI team built a staffing app that recommends people for projects and explains how their experience fits; staffing leads review a shortlist with supporting evidence.

Mistral says Studio deployments can run in a customer’s cloud, VPC or on-premises environment, on Mistral-hosted infrastructure, or across a combination; sensitive data and related processes can remain within the customer’s perimeter. OpenTelemetry traces record model calls, tool use and approvals, while evaluations let teams test prompt, model or tool changes against their own datasets before release. Mistral says customers use systems built with its Solutions team for tasks from penetration testing to KYC onboarding, keeping sensitive data and specialist knowledge under their control.