# Quick Start Guide
## 🚀 Get Started in 5 Minutes
### 1. Clone and Install
```bash
git clone https://github.com/yourusername/log-analyzer-mcp.git
cd log-analyzer-mcp
pip install -r requirements.txt
```
### 2. Configure AWS Credentials
```bash
cp .env.example .env
```
Edit `.env`:
```env
AWS_ACCESS_KEY_ID=your_access_key_here
AWS_SECRET_ACCESS_KEY=your_secret_key_here
AWS_REGION=us-east-2
```
### 3. Test the Server
```bash
python server.py
```
If you see "Starting Log Analyzer MCP Server..." - you're ready! Press Ctrl+C to stop.
### 4. Use with Claude Desktop
#### Find Your Config File
- **macOS**: `~/Library/Application Support/Claude/claude_desktop_config.json`
- **Windows**: `%APPDATA%\Claude\claude_desktop_config.json`
- **Linux**: `~/.config/Claude/claude_desktop_config.json`
#### Add This Configuration
```json
{
"mcpServers": {
"log-analyzer": {
"command": "python",
"args": ["/full/path/to/log-analyzer-mcp/server.py"],
"env": {
"AWS_ACCESS_KEY_ID": "your_key",
"AWS_SECRET_ACCESS_KEY": "your_secret",
"AWS_REGION": "us-east-2"
}
}
}
}
```
**Important**: Use the FULL absolute path to `server.py`
### 5. Restart Claude Desktop
Completely quit and restart Claude Desktop.
### 6. Try It Out!
In Claude Desktop, try these commands:
```
1. Process some logs:
"Use fetch_local_logs to process logs from /path/to/my/logs"
2. Vectorize them:
"Use store_chunks_as_vectors to create embeddings"
3. Query them:
"Use query_SFlogs to find all NullPointerExceptions"
```
## 📁 Example Workflow
### Step 1: Prepare Your Logs
```bash
# Create a logs folder
mkdir logs
# Copy your log files there
cp /path/to/app.log logs/
cp /path/to/error.log logs/
```
### Step 2: Process Logs
In Claude:
```
Use fetch_local_logs with input_folder="./logs", chunk_size=4096, overlap=1024
```
### Step 3: Vectorize
In Claude:
```
Use store_chunks_as_vectors with use_cache=true
```
### Step 4: Query
In Claude:
```
Query logs: "What errors occurred today?"
Query logs: "Summarize all exceptions"
Query logs: "Find timeouts in UserService"
```
## 🔍 Common Use Cases
### Find All Errors
```
Use query_SFlogs with query="list all errors"
```
### Root Cause Analysis
```
Use query_SFlogs with query="what caused the NullPointerException?"
```
### Time-based Analysis
```
Use query_SFlogs with query="what happened between 10am and 11am?"
```
### Class-specific Issues
```
Use query_SFlogs with query="show errors in CustomerController"
```
## ⚡ Pro Tips
1. **First Time**: Process → Vectorize → Query (in that order)
2. **Caching**: Vectorization is cached - reprocess is instant!
3. **Large Logs**: Use smaller chunk_size (e.g., 2048) for better precision
4. **Specific Searches**: Include class names or error types in queries
5. **Summaries**: Ask for "summary" or "overview" for comprehensive analysis
## 🛠️ Troubleshooting
### "No vector JSON found"
→ Run `store_chunks_as_vectors` first
### "Input folder not found"
→ Use absolute paths or check current directory
### "AWS credentials not found"
→ Verify `.env` file has correct credentials
### MCP server not responding
→ Check Claude Desktop logs and restart the app
## 📚 Learn More
- Full documentation: [README.md](README.md)
- Deployment guide: [DEPLOYMENT.md](DEPLOYMENT.md)
- GitHub issues: Report problems or request features
## 💡 Example Session
```
You: "I have some application logs I need to analyze"
Claude: "I can help! First, let me process your logs."
[Uses fetch_local_logs]
You: "Now find all the errors"
Claude: "Let me vectorize the logs first, then search."
[Uses store_chunks_as_vectors]
[Uses query_SFlogs with query="all errors"]
Claude: "I found 15 unique error patterns:
1. NullPointerException (12 occurrences)
2. TimeoutException (8 occurrences)
..."
```
That's it! You're ready to analyze logs with AI! 🎉
