LaunchDarkly MCP Server

An MCP (Model Context Protocol) server that lets AI assistants (Claude, Cursor, etc.) manage LaunchDarkly feature flags via natural language.

Quick Start

Prerequisites

• Python 3.10+

• A LaunchDarkly API access token (create one here)

Install & Run

# Install from PyPI
pip install launchdarkly-mcp-server

# Set your API key
export LD_API_KEY="your-api-key-here"

# Run the server
launchdarkly-mcp

Or use uvx (no install needed):

export LD_API_KEY="your-api-key-here"
uvx launchdarkly-mcp-server

Claude Desktop Configuration

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "launchdarkly": {
      "command": "uvx",
      "args": ["launchdarkly-mcp-server"],
      "env": {
        "LD_API_KEY": "your-api-key-here"
      }
    }
  }
}

Restart Claude Desktop and look for the hammer icon — your LaunchDarkly tools are ready.

Related MCP server: Flipt MCP Server

Available Tools

Phase 1 — Read-only

Phase 2 — Write operations

Examples

Ask Claude:

"What projects do I have in LaunchDarkly?"

"List my feature flags in the production environment"

"Turn on the new-checkout flag in staging"

"Who made changes to the dark-mode flag in the last week?"

"Add user alice@example.com to the beta flag targeting"

Development

git clone https://github.com/your-org/launchdarkly-mcp-server
cd launchdarkly-mcp-server
uv venv
uv pip install -e ".[dev]"

Run tests:

uv run pytest

Architecture

AI Assistant ── MCP (stdio) ──► launchdarkly-mcp-server ── HTTPS ──► LaunchDarkly API

Built with fastmcp (Python).

License

MIT