# ⚙️ Configuration Reference
Complete guide for configuring your Article Quadrant Analyzer MCP server.
## 📋 Table of Contents
- [Environment Variables](#environment-variables)
- [IDE Configurations](#ide-configurations)
- [MCP Server Settings](#mcp-server-settings)
- [Content Processing Limits](#content-processing-limits)
## 🔧 Environment Variables
Create a `.env` file (copy from `.env.example`):
```bash
# Required: Mistral Document AI API for OCR functionality
MISTRAL_API_KEY=your_mistral_api_key_here
# Optional: Content processing limits
CONTENT_MAX_LENGTH=50000
OCR_MAX_FILE_SIZE=10485760
# Optional: Logging level (DEBUG, INFO, WARNING, ERROR)
LOG_LEVEL=INFO
# Optional: HTTP timeout for content extraction
HTTP_TIMEOUT=30
```
### Getting Mistral API Key
1. Visit [Mistral AI Console](https://console.mistral.ai/)
2. Create an account or sign in
3. Navigate to API Keys section
4. Generate a new API key
5. Add to your `.env` file
## 💻 IDE Configurations
### Cursor
**Location**: `~/.cursor/mcp_settings.json`
```json
{
"mcpServers": {
"article-quadrant-analyzer": {
"command": "uvx",
"args": [
"--quiet",
"--python",
"3.12",
"--with",
"fastmcp",
"python",
"/absolute/path/to/test_simple_server.py"
],
"env": {
"LOG_LEVEL": "INFO"
}
}
}
}
```
### VS Code
**Location**: `~/.vscode/mcp_settings.json`
```json
{
"mcpServers": {
"article-quadrant-analyzer": {
"command": "uvx",
"args": [
"--quiet",
"--python",
"3.12",
"--with",
"fastmcp",
"python",
"/absolute/path/to/test_simple_server.py"
],
"env": {
"LOG_LEVEL": "INFO"
}
}
}
}
```
### Claude Desktop
**Location**: `~/Library/Application Support/Claude/claude_desktop_config.json`
```json
{
"mcpServers": {
"article-quadrant-analyzer": {
"command": "uvx",
"args": [
"--quiet",
"--python",
"3.12",
"--with",
"fastmcp",
"python",
"/absolute/path/to/test_simple_server.py"
],
"env": {
"LOG_LEVEL": "INFO"
}
}
}
}
```
### Emacs
**Location**: Add to your Emacs config
```elisp
;; In your init.el or config file
(with-eval-after-load 'mcp
(mcp-register-server
"article-quadrant-analyzer"
:command "uvx"
:args '("--quiet" "--python" "3.12" "--with" "fastmcp" "python" "/absolute/path/to/test_simple_server.py")
:env '(("LOG_LEVEL" . "INFO"))))
```
### Neovim
**Location**: Add to your Neovim config
```lua
-- In your init.lua
require('mcp').setup({
servers = {
['article-quadrant-analyzer'] = {
command = 'uvx',
args = {'--quiet', '--python', '3.12', '--with', 'fastmcp', 'python', '/absolute/path/to/test_simple_server.py'},
env = { LOG_LEVEL = 'INFO' }
}
}
})
```
## ⚙️ MCP Server Settings
### Transport Protocols
The server supports both stdio and HTTP transport:
**Stdio (Recommended for IDEs)**:
- Default configuration
- Better performance for IDE integration
- Simpler setup
**HTTP (For web applications)**:
```bash
# Start HTTP server
fastmcp dev --transport http test_simple_server.py
# Runs on http://127.0.0.1:PORT
```
### UVX Configuration
UVX provides zero-dependency deployment:
```bash
# Standard UVX command
uvx --quiet --python 3.12 --with fastmcp python test_simple_server.py
# With specific FastMCP version
uvx --quiet --python 3.12 --with "fastmcp==2.13.0.2" python test_simple_server.py
```
## 📊 Content Processing Limits
### Default Limits
```python
# Maximum content length for text processing
CONTENT_MAX_LENGTH = 50000 # characters
# Maximum file size for OCR processing
OCR_MAX_FILE_SIZE = 10485760 # bytes (10MB)
# HTTP request timeout
HTTP_TIMEOUT = 30 # seconds
```
### Adjusting Limits
Modify these in your environment or code:
```bash
# For large documents
CONTENT_MAX_LENGTH=100000
# For high-resolution images
OCR_MAX_FILE_SIZE=20971520 # 20MB
# For slow websites
HTTP_TIMEOUT=60
```
## 🔍 Validation
### Test Configuration
```bash
# Validate MCP setup
./deploy_to_ide_standard.sh validate
# Test with MCP Inspector
fastmcp dev test_simple_server.py
# Opens: http://127.0.0.1:6274
```
### Common Issues
**Path Issues**:
- Use absolute paths in configuration
- Verify the `test_simple_server.py` file exists
- Check directory permissions
**Dependency Issues**:
```bash
# Install FastMCP
pip install fastmcp
# Install with UVX (recommended)
pip install uvx
# Verify installation
fastmcp --version
```
**API Key Issues**:
- Verify Mistral API key is valid
- Check `.env` file is in project root
- Ensure API key has required permissions
## 📝 Configuration Templates
Ready-to-use templates are available in the `config/` directory:
- `config_cursor_standard.json`
- `config_vscode_standard.json`
- `config_claude_desktop_standard.json`
- `config_emacs.el`
- `config_neovim.lua`
## 🚀 Advanced Configuration
### Custom Server Settings
Modify `test_simple_server.py` for advanced configuration:
```python
# Custom server name and description
@mcp.server("custom-name")
async def custom_server():
pass
# Custom tool configurations
@mcp.tool
async def custom_tool(
ctx: Context,
input_param: str = Field(description="Custom parameter")
) -> str:
# Custom implementation
pass
```
### Performance Optimization
```bash
# Use UVX for faster startup
uvx --quiet --python 3.12 --with fastmcp python test_simple_server.py
# Enable caching
export CONTENT_CACHE_ENABLED=true
# Use HTTP transport for web apps
fastmcp dev --transport http test_simple_server.py
```
---
**📚 Need more help?**
- Check [README.md](README.md) for overview
- Try [QUICK_START.md](QUICK_START.md) for fast setup
- Review technical specs in `.trae/specs/`
