Flutter MCP Server

🚀 Flutter MCP Server

A bridge between AI assistants, development tools, and automated workflows in the Dart and Flutter ecosystem

⚠️ Important Note

Smithery's cloud deployment is not yet complete. This MCP server currently only supports local use. Smithery integration and registry list functionality are not yet available.

🚀 Quick Start

Overview

Flutter MCP Server is an open - source implementation of the Model Context Protocol (MCP) in the Dart and Flutter ecosystems. It serves as a bridge between AI assistants, development tools, and automated workflows, enabling the automation of Dart and Flutter development tasks in a secure, programmatic, and context - aware manner.

✨ Features

Functionality of the MCP Server

• Expose Dart & Flutter SDK tools (analyze, format, fix, create, run, test) as programmable endpoints, allowing AI agents, scripts, or remote clients to call them via the MCP protocol.
• Provide a unified API for operations such as code analysis, formatting, project building, and running tests - remotely and securely.
• Securely manage environment variables using best practices to ensure that secrets and configurations are not exposed.
• Support resource endpoints for document queries, news, and community examples, enabling AI clients to easily retrieve relevant information for Dart/Flutter projects.
• Designed for platform integration such as Smithery, but can also be run locally or deployed in any Docker - compatible environment.

Problems Solved

• Bridge the gap between AI and code: Allow AI - driven tools (such as code assistants, chatbots, or CI agents) to interact with real Dart/Flutter projects, call tools, and perform context - aware programmatic operations.
• Automate repetitive tasks: Developers and teams can automatically perform operations such as code formatting, analysis, testing, and project creation through a unified and secure interface.
• Standardize tool calls: Eliminate the need for custom scripts or manual tool execution - all content is exposed through a consistent and discoverable API.
• Support remote and cloud workflows: Can be deployed in CI/CD, cloud platforms, or AI - driven development pipelines.
• Promote best practices: Handle secrets through environment variables, support containerized deployment, and follow open standards for protocol and API design.

Feature Completeness

• Tool endpoints: Invocation of Dart and Flutter tools such as analyze, format, fix, create, run, test.
• Resource management: Access and query of resources such as documentation, sample code, and news.
• Environment variable support: Configure server behavior and functionality through environment variables.
• Scalability: Support custom plugins and extensions to meet specific project requirements.

📦 Installation

Local Installation

1. Clone the repository:

   git clone https://github.com/your-repository/flutter-mcp-server.git
   cd flutter-mcp-server
   

2. Create a virtual environment (recommended) and install dependencies:

   python -m venv env
   source env/bin/activate  # Use `env\Scripts\activate` on Windows
   pip install -r requirements.txt
   

3. Start the server:

   python server.py
   

Deployment (Smithery)

1. Associate the project repository with your Smithery account.
2. Push code changes to trigger a build.
3. Configure environment variables and deployment parameters in the Smithery interface.
4. Use the Playground or a compatible client provided by Smithery to verify the server's running status.

📚 Documentation

Server Settings

Edit the config.json file:

{
  "host": "localhost",
  "port": 5000,
  "debug_mode": true,
  "api_key": "your_api_key_here"
}

Environment Variables

Set the following environment variables during deployment or runtime:

• FLUTTER_PATH: Specify the Flutter installation path.
• DART-sdk_PATH: Specify the Dart SDK installation path.
• MCP_SERVER_MODE: Set it to development or production.

💻 Usage Examples

Basic Usage

CLI Invocation

# Analyze code
mcp-server analyze --path ./my_project/

# Format code
mcp-server format --path ./my_project/ --fix

# Create a new project
mcp-server create-project --name my_new_project --type flutter

API Invocation

Use an HTTP request:

curl -X POST http://localhost:5000/api/analyze \
     -H "Content-Type: application/json" \
     -d '{"path":"./my_project/"}'

📄 License

MIT