# LaunchDarkly MCP > LaunchDarkly MCP connects your AI client directly to feature flag systems, environments, and experiment data. You manage complex deployments by simply asking questions—checking if a new button is visible only in beta users' accounts or tracking how many people are engaging with the latest UI change. ## Overview - **Category:** ship-it - **Price:** Free - **Tags:** feature-flags, deployment-strategy, experimentation, release-management, environment-variables, ci-cd-integration ## Description Feature flags complicate releases; they’re what let developers roll out code piece-by-piece instead of flipping everything on at once. This MCP lets you talk to your LaunchDarkly platform and manage all those complex rules conversationally. You can check specific feature flag statuses across multiple environments, inspect which projects are active, or even pull historical audit logs to see who changed what and when. Need to know if the new payment flow is live in Production but only for 10% of users? Your agent handles that query instantly. It's all managed through Vinkius, giving your AI client one place to access crucial deployment intelligence. You simply ask your agent to list available environments or check specific metrics without ever touching a dashboard. ## Tools ### get_environment Fetches specific operational details about one particular workspace or environment. ### get_feature_flag Provides deep, detailed information about one single feature flag by its name or ID. ### get_metric Retrieves detailed data for a specific metric, helping you evaluate test performance. ### get_project Fetches general configuration details for a specific LaunchDarkly project. ### list_audit_logs Retrieves an account-wide history of changes, showing who modified settings and when. ### list_environments Pulls every active environment (like Dev, Staging, or Production) within the current project. ### list_feature_flags Retrieves a comprehensive list of all feature flags defined in a specific project. ### list_metrics Gets a list of all available experimentation metrics within the current project scope. ### list_projects Retrieves a list of all available LaunchDarkly projects attached to your account. ## Prompt Examples **Prompt:** ``` Check if the newly implemented dark mode feature flag is switched on in Production. ``` **Response:** ``` Searching the Production workspace... The UI Dark Mode feature toggle is currently set to ON targeting 50% of the active user base. ``` **Prompt:** ``` Turn off the experimental flag targeting our staging environment immediately. ``` **Response:** ``` Executing shutdown on experimental flag. The flag has been returned to its OFF (kill switch) state in Staging. ``` **Prompt:** ``` List all active environments linked to our main workspace project. ``` **Response:** ``` Scanning project environments... I found 3 connected clusters: Staging, Production, and Beta Testing Node. ``` ## Capabilities ### Check flag status across multiple environments The MCP retrieves the current on/off state of any feature flag within specified workspaces, like staging or production. ### List and inspect all project configurations You can pull a list of every LaunchDarkly project associated with your account to understand your deployment scope. ### Examine experiment tracking metrics The system gathers data on specific user engagement metrics, allowing you to evaluate A/B test performance directly from the chat. ### Identify all active workspaces It lists every environment (development, staging, production) connected to your main account for comprehensive oversight. ### View historical deployment changes You can retrieve full audit log entries detailing who made changes and when they happened across the entire platform. ## Use Cases ### The Production Flag Check A Product Manager needs to know if the new checkout UI flag is active for beta users in Production. They ask their agent: 'Check the status of the checkout-beta-ui flag.' The MCP replies with the exact rollout percentage and environment details, saving them a 15-minute manual check. ### The Post-Incident Investigation A DevOps Engineer notices a bug and needs to know when the problematic configuration was deployed. They call list_audit_logs, retrieving a timeline that shows exactly which user changed the setting and what time it happened. ### The Environment Map A Fullstack Developer joins a new project and needs to know all available deployment targets. They ask for 'all connected workspaces,' triggering list_environments, instantly mapping out Staging, QA, and Beta nodes. ### Comparing Feature Performance The team is debating two different onboarding flows (A vs B). Instead of manually pulling graphs, they ask the agent to check 'onboarding-flow-metrics,' getting real-time data via list_metrics. ## Benefits - Cut out the dashboard hopping. Instead of clicking through multiple tabs to check a feature status, you just ask your agent and get the answer immediately, whether it's in Staging or Production. - Audit every change without searching logs for hours. Use list_audit_logs to pull an entire history, letting you see who made which flag modification and exactly when it happened. - Evaluate A/B tests right away. By calling list_metrics and get_metric, you stop guessing about user behavior and start basing decisions on real data. - Know your boundaries. You can use list_projects to map out every single deployment silo, making sure nothing gets accidentally missed during a release cycle. - Get immediate status checks. Whether you need the current state of a feature flag via get_feature_flag or just want to see all connected workspaces using list_environments, it's instant. ## How It Works The bottom line is you stop clicking through dashboards and start asking questions about your feature flags and deployments. 1. Install this MCP locally and supply your LaunchDarkly API token key. 2. Connect your AI client to the Vinkius catalog using the credentials. 3. Use natural language commands to request specific flag statuses, environment lists, or project metrics. ## Frequently Asked Questions **How do I list all available deployments using LaunchDarkly MCP?** You use list_projects to get an overview of all connected projects. Then, call list_environments to see the specific deployment targets (like Staging or Production) for that project. **Can I check flag status without knowing the exact flag name with LaunchDarkly MCP?** First, use list_feature_flags to retrieve a comprehensive list of all available flags. Then, you can provide the specific name or ID when calling get_feature_flag. **Does LaunchDarkly MCP help with debugging who changed what?** Yes, running list_audit_logs retrieves a full history of every change made across your account. This is crucial for understanding the sequence of events that led to an issue. **Is LaunchDarkly MCP only useful for Production environments?** No, it works with all environments. You can use list_environments and get_environment to check status in Dev, Staging, or any other workspace you need to validate. **How does LaunchDarkly MCP help me evaluate A/B tests?** You call list_metrics and then get_metric. This pulls specific user engagement data for the test flags, letting you see if one version is performing better than another.