Mistral AI has released managed agents, apps, and workflow tools for its Studio platform, with general availability for workflows and public previews for the other features now accessible on Mistral Cloud.
The release expands Mistral Studio’s capabilities to include structured automation alongside autonomous agent execution. According to the announcement page1, workflows are generally available today, allowing users to combine fixed rules with agent reasoning and human approvals. Managed agents and apps are currently in public preview. These new features integrate with Mistral Large 4 (Le Chonk) and existing enterprise data connectors.
Workflows define process logic, enabling teams to automate steps that require both rule-based execution and agent-driven decision-making. The platform allows users to describe processes in Mistral Vibe to generate, test, and deploy workflows, with the ability to resume interrupted runs. Managed agents allow developers to build systems that investigate and act on complex tasks, equipped with skills, sandboxes, subagents, and hooks. Studio manages the runtime infrastructure, handling hosting and execution while allowing developers to set boundaries on tool access and data usage. Apps provide a user interface for employees to interact with these systems, featuring built-in hosting, authentication, and access controls.
Availability varies by feature. Workflows are generally available on Mistral Cloud immediately. Managed agents and apps are in public preview on the same cloud environment. Private cloud and on-premises deployments for all three features are scheduled for late November, according to the announcement.
Mistral provided specific examples of internal efficiency gains resulting from the new Studio features. The company stated that its product operations team uses a workflow to route feedback and ideas, saving approximately 35 hours per week. The engineering team utilizes a managed agent to investigate production incidents, which the company claims saves hundreds of hours each month. The Applied AI team developed a staffing app that matches personnel to projects, reportedly helping assemble teams faster. These figures are vendor claims based on internal usage. The announcement includes no independent test results verifying the time-savings metrics or the performance of the new agent capabilities.
The platform introduces observability features intended to help enterprises monitor agent behavior. Studio provides OpenTelemetry traces that show each step of a run, including model calls, tool use, and human approvals. Evaluations can be run against proprietary datasets before release to test changes to prompts, models, or tools. Access controls and human approval gates are integrated into the workflow design. Developers can specify which actions require human sign-off, ensuring that sensitive operations do not proceed without oversight. The platform supports the Model Context Protocol (MCP), allowing developers to connect external data sources and tools to the agents and workflows.
The announcement notes that customers such as ABANCA, Accenture, BNP Paribas, Capgemini, CMA CGM, and Tata Consultancy Services already use Studio in production.