GrowthBook MCP Server

{ "manifest_version": "0.2", "name": "GrowthBook", "version": "1.2.0", "description": "MCP Server for interacting with GrowthBook - an open source feature flagging and experimentation platform", "long_description": "The GrowthBook MCP Server enables AI assistants to interact with GrowthBook, an open source feature flagging and experimentation platform. Create feature flags, set up A/B tests, analyze experiment results, and manage your experimentation workflow directly from your AI coding assistant.\n\n## Usage Examples\n\nAsk Claude to:\n1. Create a feature flag called `new-checkout` that defaults to `false`\n2. Summarize my last 20 experiments\n3. Create an A/B test comparing two checkout button colors - control with `#007bff` and variation with `#28a745`\n4. Search the GrowthBook docs for how to set up a fact metric", "author": { "name": "Ryan Feigenbaum", "email": "ryan@growthbook.io" }, "homepage": "https://growthbook.io", "documentation": "https://docs.growthbook.io/integrations/mcp", "icon": "icon.png", "server": { "type": "node", "entry_point": "server/index.js", "mcp_config": { "command": "node", "args": ["${__dirname}/server/index.js"], "env": { "GB_API_KEY": "${user_config.GB_API_KEY}", "GB_EMAIL": "${user_config.GB_EMAIL}", "GB_API_URL": "${user_config.GB_API_URL}", "GB_APP_ORIGIN": "${user_config.GB_APP_ORIGIN}" } } }, "tools": [ { "name": "get_environments", "description": "List all environments (e.g., production, staging) configured in GrowthBook." }, { "name": "get_projects", "description": "List all projects in your GrowthBook instance." }, { "name": "get_sdk_connections", "description": "List SDK connections (how GrowthBook connects to your apps)." }, { "name": "create_sdk_connection", "description": "Create a new SDK connection for your app, specifying language and environment." }, { "name": "create_feature_flag", "description": "Create, add, or wrap an element with a feature flag. Accepts key, type, default value, and other metadata." }, { "name": "create_force_rule", "description": "Create a feature flag with a targeting condition (e.g., `country = CA`)." }, { "name": "get_feature_flags", "description": "List feature flags in your GrowthBook instance." }, { "name": "get_single_feature_flag", "description": "Fetch details for a specific feature flag by ID." }, { "name": "get_stale_safe_rollouts", "description": "List safe rollout rules that have been rolled back or released. This tools is helpful for removing stale flags from your codebase." }, { "name": "generate_flag_types", "description": "Generate TypeScript types for feature flags." }, { "name": "get_experiments", "description": "List experiments in GrowthBook." }, { "name": "get_experiment", "description": "Fetch details for a specific experiment by ID." }, { "name": "get_attributes", "description": "List user attributes tracked in GrowthBook (useful for targeting)." }, { "name": "create_experiment", "description": "Creates a feature-flag linked experiment. Note that this tool calls `get_defaults`, which provides a sampling of previous experiments to help the AI agent create the experiment correctly. For better performance, these defaults are saved to the local file system and periodically updated." }, { "name": "get_metrics", "description": "List fact and legacy metrics (useful for better understanding experiments)." }, { "name": "get_metric", "description": "Fetch a specific fact or legacy metric by ID." }, { "name": "get_defaults", "description": "Get default values for experiments including hypothesis, description, datasource, and assignment query. (Runs automatically when the create experiment tool is called.)" }, { "name": "set_user_defaults", "description": "Set custom default values for experiments that will be used when creating new experiments." }, { "name": "clear_user_defaults", "description": "Clear user-defined defaults and revert to automatic defaults." }, { "name": "search_growthbook_docs", "description": "Search the GrowthBook documentation for information on how to use a feature, by keyword or question." } ], "prompts": [ { "name": "experiment-analysis", "description": "Analyze recent experiments and give me actionable advice", "text": "Use GrowthBook to fetch my recent experiments. Analyze the experiments and tell me:\n\n1. Which experiment types are actually worth running vs. theater?\n\n2. What's the one pattern in our losses that we're blind to?\n\n3. If you could only run 3 experiments next quarter based on these results, what would they be and why?\n\n4. What's the biggest methodological risk in our current approach that could be invalidating results?\n\nBe specific. Use the actual data. Don't give me generic advice." } ], "user_config": { "GB_API_KEY": { "type": "string", "title": "GrowthBook API key or PAT", "description": "A GrowthBook API key or PAT. When using a PAT, you can only access tools that your permissions allow.", "required": true, "sensitive": true }, "GB_EMAIL": { "type": "string", "title": "GrowthBook email", "description": "Your email address used with GrowthBook. Used when creating feature flags and experiments.", "required": true, "sensitive": false }, "GB_API_URL": { "type": "string", "title": "API URL", "description": "Your GrowthBook API URL.", "required": false, "sensitive": false, "default": "https://api.growthbook.io" }, "GB_APP_ORIGIN": { "type": "string", "title": "GrowthBook app origin URL", "description": "Your GrowthBook app URL.", "required": false, "sensitive": false, "default": "https://app.growthbook.io" } }, "keywords": ["experimentation", "feature flags", "ab testing", "growthbook"], "license": "MIT", "repository": { "type": "git", "url": "https://github.com/growthbook/growthbook-mcp" } }