# MCP Configuration Guide
## Quick Setup for MCP Clients
This guide helps you configure the Layer.ai MCP Server with various MCP-compatible clients for seamless AI asset generation.
## Prerequisites
1. **MCP-compatible client** (Claude Desktop, etc.)
2. **Python 3.10+** available in your system PATH
3. **Layer.ai account** with API access
## Installation
### 1. Install the Server
```bash
git clone https://github.com/bahadirbklg/layer-ai-mcp-server.git
cd layer-ai-mcp-server
./install.sh
```
### 2. Set Up Credentials
```bash
python layer-mcp-server/setup.py
```
### 3. Configure Your MCP Client
#### For Claude Desktop
**macOS**: `~/Library/Application Support/Claude/claude_desktop_config.json`
**Windows**: `%APPDATA%/Claude/claude_desktop_config.json`
```json
{
"mcpServers": {
"layer-ai": {
"command": "python",
"args": ["path/to/layer-mcp-server/server.py"],
"env": {
"LAYER_API_TOKEN": "pat_your_token_here",
"LAYER_WORKSPACE_ID": "your_workspace_id"
}
}
}
}
```
#### For Other MCP Clients
Use this general configuration pattern:
```json
{
"name": "layer-ai",
"command": "python",
"args": ["path/to/layer-mcp-server/server.py"],
"environment": {
"LAYER_API_TOKEN": "pat_your_token_here",
"LAYER_WORKSPACE_ID": "your_workspace_id"
}
}
```
## Usage
### Available Tools
1. **create_asset** - Generate 2D game assets
2. **get_workspace_info** - Check workspace status
3. **remove_background** - AI background removal (in development)
4. **describe_image** - AI image analysis (in development)
5. **generate_prompt** - Prompt optimization (in development)
### Example Usage
```
Generate a pixel art character for my platformer game
```
The MCP server will:
1. Process your request
2. Generate the asset using Layer.ai
3. Save it to your assets directory
4. Provide the file path and details
## Troubleshooting
### Common Issues
**"Server not found"**
- Check the file path in your MCP configuration
- Ensure Python is in your system PATH
- Verify the server.py file exists
**"Authentication failed"**
- Run the credential setup: `python layer-mcp-server/setup.py`
- Check your API token at https://app.layer.ai/settings/api-keys
- Verify workspace ID in Layer.ai dashboard
**"Tool not found"**
- Restart your MCP client
- Check server logs for errors
- Verify the server started successfully
**"Generation failed"**
- Check your Layer.ai quota (600 assets on free tier)
- Verify internet connection
- Check Layer.ai service status
### Restart MCP Server
1. Close your MCP client
2. Restart the client application
3. The server will reconnect automatically
### Configuration Validation
```bash
# Test server connection
python layer-mcp-server/server.py --test
# Verify credentials
python layer-mcp-server/setup.py --verify
```
## Advanced Configuration
### Environment Variables
You can set these in your MCP client configuration:
- `LAYER_API_TOKEN` - Your Layer.ai API token (required)
- `LAYER_WORKSPACE_ID` - Your workspace ID (required)
- `LAYER_API_BASE_URL` - API endpoint (optional)
- `LAYER_DEFAULT_SAVE_DIR` - Asset save directory (optional)
### Timeout Settings
For complex asset generation, increase timeout in your MCP client:
```json
{
"mcpServers": {
"layer-ai": {
"command": "python",
"args": ["path/to/layer-mcp-server/server.py"],
"timeout": 180,
"env": {
"LAYER_API_TOKEN": "pat_your_token_here"
}
}
}
}
```
## Security
- API tokens are encrypted and stored locally
- No credentials are sent to version control
- Secure file permissions (600) on credential files
- Local-only credential storage
## Additional Resources
- [MCP Protocol Specification](https://modelcontextprotocol.io)
- [Layer.ai API Documentation](https://docs.layer.ai)
- [Layer.ai Platform](https://app.layer.ai)
