Local Logs MCP Server
# Local Logs MCP Server
A **Model Context Protocol (MCP) server** for monitoring local application logs with real-time tailing, error tracking, and log search capabilities.
Perfect for monitoring Node.js applications, web servers, or any application that writes to log files.


## ⨠Features
- š **Log File Discovery** - Automatically finds and lists available log files
- š **Real-time Log Tailing** - Get the last N lines from any log file
- ā ļø **Error Monitoring** - Quickly check error logs for issues
- š **Server Status** - Get server status summary from log analysis
- š **Log Watching** - Monitor log files for changes
- š **Log Search** - Search for specific text patterns in logs
- š§ **Configurable** - Supports custom log directories and file extensions
- š **Easy Setup** - Works with Cursor, Claude Desktop, VS Code Copilot, and more
## š ļø Available Tools
| Tool | Description | Parameters |
|------|-------------|------------|
| `get_log_files` | List available log files with metadata | None |
| `tail_log` | Get last N lines from a log file | `filename`, `lines` |
| `get_errors` | Get recent error log entries | `lines` |
| `get_server_status` | Server status summary from logs | None |
| `watch_log` | Monitor log file for changes | `filename` |
| `search_logs` | Search for text in log files | `query`, `filename`, `lines` |
## š¦ Installation
### š One-Command Setup (Recommended)
**Linux/Mac:**
```bash
curl -sSL https://raw.githubusercontent.com/mariosss/local-logs-mcp-server/main/install-new.js | node
```
**Windows (PowerShell):**
```powershell
Invoke-WebRequest -Uri "https://raw.githubusercontent.com/mariosss/local-logs-mcp-server/main/install-new.js" | Invoke-Expression
```
**That's it!** The installation script will download the server file directly and configure Cursor automatically. Just restart Cursor and you're ready to go!
### Alternative: Two-Command Setup
```bash
npm install -g https://github.com/mariosss/local-logs-mcp-server
local-logs-setup
```
### Alternative: Manual Setup
```bash
npm install -g local-logs-mcp-server
local-logs-setup
```
### NPX Usage (No Installation)
```bash
npx local-logs-mcp-server
```
### Manual Installation
```bash
git clone https://github.com/mariosss/local-logs-mcp-server.git
cd local-logs-mcp-server
npm install -g .
```
## āļø Configuration
### ā
Automatic Configuration (Default)
The package automatically configures Cursor for you! No manual setup needed.
### š§ Manual Configuration (If needed)
#### For Cursor IDE
Add to your `.cursor/mcp.json` (project-specific) or global MCP config:
```json
{
"mcpServers": {
"local-logs": {
"command": "npx",
"args": ["-y", "local-logs-mcp-server"],
"env": {
"LOGS_DIR": "./logs"
}
}
}
}
```
### For Claude Desktop
Add to your Claude Desktop MCP config (`~/Library/Application Support/Claude/claude_desktop_config.json` on Mac):
```json
{
"mcpServers": {
"local-logs": {
"command": "npx",
"args": ["-y", "local-logs-mcp-server"],
"env": {
"LOGS_DIR": "/path/to/your/logs"
}
}
}
}
```
### For VS Code (GitHub Copilot)
Add to your `.vscode/mcp.json`:
```json
{
"servers": {
"local-logs": {
"command": "npx",
"args": ["-y", "local-logs-mcp-server"],
"env": {
"LOGS_DIR": "./logs"
}
}
}
}
```
### For Windsurf (Codium)
Add to your Windsurf MCP config:
```json
{
"mcpServers": {
"local-logs": {
"command": "npx",
"args": ["-y", "local-logs-mcp-server"],
"env": {
"LOGS_DIR": "./logs"
}
}
}
}
```
## š§ Environment Variables
| Variable | Description | Default |
|----------|-------------|---------|
| `LOGS_DIR` | Directory containing log files | Auto-detected (see below) |
| `LOG_EXTENSIONS` | Comma-separated log file extensions | `.log,.txt` |
### Log Directory Auto-Detection
If `LOGS_DIR` is not specified, the server automatically searches for logs in this priority order:
1. `./logs` (current directory)
2. `./apps/backend/logs` (monorepo structure)
3. `./server/logs` (server directory)
4. `./backend/logs` (backend directory)
5. `/var/log` (system logs on Linux/Mac)
6. `C:\logs` (system logs on Windows)
## š Usage Examples
Once configured with your MCP client, you can use natural language commands:
### Basic Commands
- *"Check my server logs"*
- *"Are there any errors in the logs?"*
- *"What's my server status?"*
- *"Show me available log files"*
### Advanced Commands
- *"Show me the last 50 lines from error.log"*
- *"Search logs for 'database connection'"*
- *"Monitor access.log for changes"*
- *"Find all log entries containing 'TypeError'"*
## šļø Project Structure Examples
### Standard Node.js Project
```
your-project/
āāā logs/
ā āāā combined.log # Main application log
ā āāā error.log # Error log
ā āāā access.log # Access log (optional)
āāā .cursor/
ā āāā mcp.json # MCP configuration
āāā src/
āāā package.json
```
### Monorepo Structure
```
your-monorepo/
āāā apps/
ā āāā backend/
ā āāā logs/
ā āāā combined.log
ā āāā error.log
āāā .cursor/
ā āāā mcp.json
āāā package.json
```
## š Supported Log Formats
The server works with any text-based log files and can parse:
- **Winston logs** (JSON and text format)
- **Morgan access logs**
- **Pino logs**
- **Bunyan logs**
- **Custom application logs**
- **System logs**
- **Any .log or .txt files**
### Example Log Formats
**Winston Text Format:**
```
2024-01-15 10:30:45 [INFO]: Server started on port 3000
2024-01-15 10:30:50 [ERROR]: Database connection failed
```
**Winston JSON Format:**
```json
{"level":"info","message":"Server started","timestamp":"2024-01-15T10:30:45.123Z"}
{"level":"error","message":"Database connection failed","timestamp":"2024-01-15T10:30:50.456Z"}
```
## š”ļø Security
- **Read-only access** - Server only reads log files, never writes or modifies
- **Directory scoping** - Only accesses files in the specified logs directory
- **No network access** - Operates entirely on local files
- **Permission respect** - Respects file system permissions
## š Performance
- **Efficient file reading** - Uses streaming for large log files
- **Memory conscious** - Doesn't load entire files into memory
- **Fast search** - Optimized text search algorithms
- **Minimal dependencies** - Only uses Node.js built-in modules
## š¤ Contributing
1. Fork the repository
2. Create your feature branch (`git checkout -b feature/amazing-feature`)
3. Commit your changes (`git commit -m 'Add amazing feature'`)
4. Push to the branch (`git push origin feature/amazing-feature`)
5. Open a Pull Request
## š License
This project is licensed under the MIT License - see the [LICENSE](LICENSE) file for details.
## š Troubleshooting
### Server shows red/error status
- Check that the logs directory exists and is accessible
- Verify log files have proper read permissions
- Ensure Node.js version is 14 or higher
- Check that the server path is correct in your MCP config
### No tools available
- Restart your MCP client (Cursor, Claude, VS Code)
- Verify MCP configuration syntax is valid JSON
- Check that the MCP server is responding (test with `echo '{"jsonrpc":"2.0","method":"initialize","id":1}' | npx local-logs-mcp-server`)
### Log files not found
- Set the correct `LOGS_DIR` environment variable
- Make sure your application is actually writing to log files
- Check that file extensions match `LOG_EXTENSIONS` setting
- Verify directory permissions allow reading
### Common Log Directory Issues
**Issue:** "Logs directory not found"
```bash
# Solution: Set explicit path
export LOGS_DIR="/path/to/your/logs"
```
**Issue:** "No log files found"
```bash
# Solution: Check file extensions
export LOG_EXTENSIONS=".log,.txt,.out"
```
## š Examples
### Complete Cursor Setup
1. **Install globally:**
```bash
npm install -g local-logs-mcp-server
```
2. **Add to `.cursor/mcp.json`:**
```json
{
"mcpServers": {
"local-logs": {
"command": "local-logs-mcp",
"env": {
"LOGS_DIR": "./logs"
}
}
}
}
```
3. **Restart Cursor and test:**
- Ask: *"Check my server logs"*
### Complete Claude Desktop Setup
1. **Open Claude Desktop config:**
```bash
# Mac
open ~/Library/Application\ Support/Claude/claude_desktop_config.json
# Windows
notepad %APPDATA%\Claude\claude_desktop_config.json
```
2. **Add configuration:**
```json
{
"mcpServers": {
"local-logs": {
"command": "npx",
"args": ["-y", "local-logs-mcp-server"],
"env": {
"LOGS_DIR": "/Users/yourusername/projects/myapp/logs"
}
}
}
}
```
3. **Restart Claude Desktop**
## š Related Projects
- [Supabase MCP Server](https://github.com/supabase/mcp-server-supabase)
- [Model Context Protocol](https://modelcontextprotocol.io/)
- [MCP Servers Collection](https://github.com/modelcontextprotocol/servers)
- [Cursor IDE](https://cursor.sh/)
- [Claude Desktop](https://claude.ai/desktop)
---
**Made with ā¤ļø for the MCP community**
TDQS
Scored across 6 tools
Each tool has a clearly distinct purpose with no overlap: get_errors retrieves error entries, get_log_files lists files, get_server_status provides a summary, search_logs finds text, tail_log gets last lines, and watch_log monitors changes. The descriptions reinforce these unique functions, making tool selection unambiguous.
All tool names follow a consistent verb_noun pattern using snake_case, such as get_errors, search_logs, and tail_log. This uniformity makes the set predictable and easy to understand, with no deviations in naming style.
With 6 tools, the server is well-scoped for log management, covering essential operations like listing, searching, tailing, and monitoring logs. Each tool earns its place without being excessive or insufficient for the domain.
The tool set provides strong coverage for log reading and monitoring, including search, tail, and watch functionalities. A minor gap exists in write operations (e.g., clearing or modifying logs), but agents can effectively work with the provided read-only tools for typical log analysis tasks.