# [WIP] Android Development MCP Server
A Model Context Protocol (MCP) server that provides Android development assistance following NowInAndroid best practices. This server can be used with any MCP-compatible AI tool (Claude, GPT, Gemini, etc.).
## Features
### Tools Available
| Tool | Description |
|------|-------------|
| `get_quick_reference` | Quick overview of patterns and available tools |
| `get_architecture_reference` | Full architecture documentation (Data, Domain, UI layers) |
| `get_compose_patterns` | Jetpack Compose UI patterns |
| `get_modularization_guide` | Module structure and dependencies |
| `get_gradle_setup` | Build configuration and convention plugins |
| `get_testing_patterns` | Testing approach with test doubles |
| `generate_feature_module` | Generate complete feature module code |
| `generate_viewmodel` | Generate ViewModel with UiState |
| `generate_repository` | Generate offline-first repository |
| `search_documentation` | Search across all documentation |
## Quick Start
### Option 1: Docker (Recommended)
```bash
# Build and run with Docker Compose
docker-compose up -d
# Or build and run manually
docker build -t android-dev-mcp .
docker run -d -p 3000:3000 --name android-mcp android-dev-mcp
```
The server will be available at:
- SSE endpoint: `http://localhost:3000/sse`
- Health check: `http://localhost:3000/health`
### Option 2: Local Development
```bash
# Install dependencies
npm install
# Build
npm run build
# Run in SSE mode (for network access)
npm start -- --sse
# Or run in stdio mode (for local CLI)
npm start
```
## Connecting AI Clients
### Claude Desktop
Add to your Claude Desktop config (`~/.config/claude/claude_desktop_config.json` on Linux/Mac or `%APPDATA%\Claude\claude_desktop_config.json` on Windows):
```json
{
"mcpServers": {
"android-dev": {
"transport": "sse",
"url": "http://localhost:3000/sse"
}
}
}
```
### Claude Code
```bash
claude mcp add android-dev --transport sse --url http://localhost:3000/sse
```
### Other MCP Clients
Configure your MCP client with:
- **Transport**: SSE
- **URL**: `http://localhost:3000/sse`
## Example Usage
Once connected, you can ask your AI assistant questions like:
- "What's the recommended architecture for an Android app?"
- "Generate a feature module for user-profile with package com.myapp"
- "Show me the ViewModel pattern from NowInAndroid"
- "How do I set up offline-first data sync?"
- "Generate a repository for the Task entity"
- "Search documentation for navigation"
## API Endpoints
| Endpoint | Method | Description |
|----------|--------|-------------|
| `/sse` | GET | SSE connection for MCP |
| `/messages` | POST | Message handling |
| `/health` | GET | Health check |
## Environment Variables
| Variable | Default | Description |
|----------|---------|-------------|
| `PORT` | 3000 | Server port |
| `MCP_TRANSPORT` | sse | Transport mode (sse or stdio) |
| `NODE_ENV` | production | Node environment |
## Project Structure
```
android-dev-mcp/
├── src/
│ └── index.ts # Main server code
├── references/ # Documentation files
│ ├── architecture.md
│ ├── compose-patterns.md
│ ├── gradle-setup.md
│ ├── modularization.md
│ └── testing.md
├── dist/ # Compiled JavaScript
├── Dockerfile
├── docker-compose.yml
├── package.json
└── tsconfig.json
```
## Extending the Server
### Adding New Tools
Edit `src/index.ts` and add a new tool:
```typescript
server.tool(
"my_new_tool",
"Description of what the tool does",
{
param1: {
type: "string",
description: "Parameter description",
},
},
async ({ param1 }) => ({
content: [{ type: "text", text: `Result for ${param1}` }],
})
);
```
### Adding New Documentation
1. Add markdown files to `references/`
2. Create a loader function in `src/index.ts`
3. Register a new tool to expose it
## License
MIT
## References
- [NowInAndroid](https://github.com/android/nowinandroid) - Google's reference Android app
- [Model Context Protocol](https://modelcontextprotocol.io/) - MCP specification
- [Android Architecture](https://developer.android.com/topic/architecture) - Official guide
