mcp

This MCP server from LaunchDarkly implements the Model Context Protocol (MCP) for interacting with LaunchDarkly AI-enabled workflows. It exposes an MCP-compatible interface that lets clients request AI-configured data, manage AI configurations, and integrate LaunchDarkly’s feature-flag context into an MCP-enabled agent. The server definition shown uses npx to run the official mcp-server package and authenticates requests with an API key obtained from LaunchDarkly. To use it, configure your MCP client to point at the LaunchDarkly MCP server, provide the API key, and then issue the standard MCP operations (such as get, list, create and update operations for AI configs and related resources) as supported by the server’s API surface. Tools available via this server include operations for managing AI configurations, their targetings, and variations, enabling seamless integration with your AI agent workflows.

Once running, you can connect your MCP client to the server using a configuration similar to the example provided. The server supports environment-variable-based token provisioning and can be accessed via the CLI or through your MCP client’s configuration, depending on your integration needs.

Prerequisites:

• Node.js and npm or yarn installed on your development machine
• Internet access to fetch the mcp-server package from npm

Installation steps:

1. Install Node.js and npm if not already installed. You can download from https://nodejs.org/
2. Install and run the MCP server using npx (no global install required):

npx -y --package @launchdarkly/mcp-server -- mcp start --api-key api-xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx

3. Alternatively, install the package locally and run via node if you prefer a local setup:

mkdir mcp-server-ld
cd mcp-server-ld
npm init -y
npm install @launchdarkly/mcp-server
node_node_path=$(which node)
node $NODE_PATH/node_modules/@launchdarkly/mcp-server/bin/mcp-server.js start --api-key api-xxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxxx

4. If you want to run using a standalone binary, follow the standalone binary installation steps described in the repository, then point your MCP client to the binary's start command with the API key.

5. Add your specific API key to the configuration in your MCP client as shown in the examples.

Tips and considerations:

• Your API key must be kept secure; do not commit it to public repositories. Use environment variables or secret managers where possible.
• If your AI client supports environment-variable substitution, you can inject tokens as shown in the README examples using MCP_LD_TOKEN mapped to LD_ACCESS_TOKEN.
• The example uses npx to fetch and run the mcp-server package on demand; for production, consider using a pinned version or a local installation to avoid network variability.
• If you need to run behind a firewall or in restricted environments, consider using a standalone binary released by the project and bypassing npm.
• Ensure the API key has the correct permissions on LaunchDarkly to access MCP server resources.
• When using Docker or other deployment options, verify the URL or SSE endpoint configured in your MCP client aligns with the server’s runtime endpoint.