Featureflow

You run an MCP server to let AI assistants interact with your Featureflow feature flag management, enabling you to view projects, environments, features, and more through natural conversations. This guide shows how to use the server, install it locally, and configure secure access so you can manage your feature flags hands-free.

How to use

Once the MCP server is configured in your environment, you can ask your AI assistant to perform common feature flag management tasks. Practical actions include listing all projects, creating and updating features or environments, cloning features, archiving or unarchiving flags, and reviewing current feature controls across environments. You can also retrieve targeting attributes, API keys, and environment details to plan changes with confidence.

How to install

Prerequisites: you need Node.js installed on your machine. Ensure you have a working development environment with npm available.

Install and run the MCP server locally using the following steps.

# Clone the MCP server repository
git clone https://github.com/featureflow/featureflow-mcp.git
cd featureflow-mcp

# Install dependencies
npm install

# Build the project
npm run build

# Run locally with your API token (replace YOUR_TOKEN)
FEATUREFLOW_API_TOKEN=api-YOUR_TOKEN npm start

Configuration

Configure the MCP server to connect to your Featureflow instance. Use the following example to run the MCP server via a local tool or client that supports MCP by providing the required token.

{
  "mcpServers": {
    "featureflow": {
      "command": "npx",
      "args": ["-y", "featureflow-mcp"],
      "env": {
        "FEATUREFLOW_API_TOKEN": "api-your-token-here"
      }
    }
  }
}

Security and notes

Keep your API token secret. Store tokens securely and rotate them periodically. When using a self-hosted Featureflow instance, you can also set the API URL in the environment to point to your own deployment.

Examples of available actions

Here are representative actions you can perform through the MCP-enabled assistant once configured, aligned with the available tools.

Troubleshooting

If you encounter connection issues, verify that the MCP server process is running and that the API token is correct. Ensure network access to the Featureflow API URL if you are using a self-hosted deployment.

Available tools

list_projects

List all projects, optionally filtered by a query

get_project

Get a specific project by ID or key

create_project

Create a new project

update_project

Update an existing project

delete_project

Delete a project

list_features

List features with optional filters

get_feature

Get a specific feature by ID or unified key

create_feature

Create a new feature flag

update_feature

Update an existing feature

clone_feature

Clone a feature with a new key

archive_feature

Archive or unarchive a feature

delete_feature

Delete a feature

get_feature_control

Get feature control settings for an environment

update_feature_control

Enable/disable features, modify rules

list_environments

List environments for a project

get_environment

Get a specific environment

create_environment

Create a new environment

update_environment

Update an existing environment

delete_environment

Delete an environment

list_targets

List targeting attributes for a project

get_target

Get a specific target by key

list_api_keys

List SDK API keys for an environment