MCP Server
Overview
Model Context Protocol (MCP) is an open protocol that standardizes how applications provide context to large language models (LLMs). The ConfigCat MCP server lets you manage feature flags and configurations via the Public Management API seamlessly from AI tools. It also enables your code editor to understand your feature flags, integrate the appropriate ConfigCat SDK into your project or even create new feature flags directly in your codebase.
Features
- Manage organizations, members and permissions
- Create and update products, configs, and environments
- Manage feature flags and settings
- Manage tags and user segments
- Connect with integrations and webhooks
- Track activity with audit logs and zombie flag (stale flag) reports
- Find code references linked to your features
- Integrate ConfigCat SDK and implement feature flags in your project
Prerequisites
- Node.js version 16 or higher installed
- Public Management API basic auth credentials for ConfigCat MCP server created
Setup
Supply your Public Management API basic auth credentials.
The MCP server's configuration includes the following environment variables:
Variable | Required | Default | Description |
---|---|---|---|
CONFIGCAT_API_USER | ☑ | – | Management API basic auth username. |
CONFIGCAT_API_PASS | ☑ | – | Management API basic auth password. |
CONFIGCAT_BASE_URL | https://api.configcat.com | Override API host (rarely needed). |
Client Configuration
- Cursor
- Visual Studio Code
- Claude Desktop
- Open Preferences → Cursor Settings → MCP & Integrations
- Click Add Custom MCP
- Add (or merge) the snippet below into your JSON settings:
{
"mcpServers": {
"ConfigCat": {
"command": "npx",
"args": ["-y", "@configcat/mcp-server"],
"env": {
"CONFIGCAT_API_USER": "YOUR_API_USER",
"CONFIGCAT_API_PASS": "YOUR_API_PASSWORD"
}
}
}
}
- Save – the server will start on demand.
- Create a
.vscode/mcp.json
file in your project root with the following content:
{
"servers": {
"ConfigCat": {
"command": "npx",
"args": ["-y", "@configcat/mcp-server"],
"env": {
"CONFIGCAT_API_USER": "YOUR_API_USER",
"CONFIGCAT_API_PASS": "YOUR_API_PASSWORD"
}
}
}
}
- Save the settings file. The MCP server should now be available in VS Code.
- Open Settings → Developer
- Click Edit Config
- In
claude_desktop_config.json
add:
{
"mcpServers": {
"ConfigCat": {
"command": "npx",
"args": ["-y", "@configcat/mcp-server"],
"env": {
"CONFIGCAT_API_USER": "YOUR_API_USER",
"CONFIGCAT_API_PASS": "YOUR_API_PASSWORD"
}
}
}
}
- Save & restart Claude Desktop.
Replace YOUR_API_USER
and YOUR_API_PASSWORD
environment variables with your Public Management API basic auth credentials.
Interaction
After you install the ConfigCat MCP server in your AI client, you can prompt your agent to create or manage your feature flags and configurations. Typically you need to click Run tool (or the equivalent option in your AI client) to execute the result.
For example, you could try asking
Create a boolean feature flag called "my_awesome_feature" in the "Backend" config
or
Turn the "my_awesome_feature" flag ON in all environments
or
Update the "my_awesome_feature” flag in dev environment so it’s only enabled for users in Canada