Steadybit MCP Server provides a Model Context Protocol (MCP) interface that allows AI agents and tools to interact with the Steadybit platform. It exposes a suite of operations for managing experiments and related resources: listing experiment designs, retrieving a specific design, listing and inspecting experiment executions with rich filtering (by experiment(s), environment, team, state, and time range) and pagination, retrieving individual executions, listing supported actions, environments, and teams, and listing experiment schedules. It also supports template exploration and creation of experiments from templates; however, creating from a template is guarded behind a capability flag CAPABILITIES_ENABLED_0=CREATE_EXPERIMENT_FROM_TEMPLATE and requires enabling in the environment. The server runs over STDIO transport, integrates with CLI tools like Claude Desktop for prompt-driven workflows, and writes logs to steadybit-mcp.log. To operate, you need a Steadybit account with an API token and an API URL. The create-from-template operation returns either the created experiment key or an error message. The README also provides setup and development details, including building via Maven, testing with the MCP inspector, and docker/native image options.