The Statsig MCP server is an implementation of the Model Context Protocol that connects your AI agent and assistants like Claude, Cursor, etc directly to your Statsig account. It provides structured and secure access to feature gates, dynamic configurations, experiments, layers, ranked variants, and event measurement, so your agent can evaluate feature access, retrieve assigned settings, rank variants, record exposures, and log events on your behalf.
- Feature gate evaluation: Have your agent check one or several feature gates for a user to determine which features should be available.
- Dynamic configuration and experiment retrieval: Let the agent evaluate dynamic configurations or experiments for a user and retrieve their assigned values and related details.
- Layer parameter evaluation: Direct your agent to evaluate a layer and fetch the parameter values assigned to a user.
- Contextual variant ranking: Instruct your agent to rank variants for a contextual bandit Autotune configuration using relevant user context.
- Exposure and event measurement: Have your agent record custom gate, experiment, or configuration exposures and log batches of events for analytics and metric calculation.