Connect Pulse to Claude Code via MCP. Find slow queries, identify bottlenecks, and get optimization suggestions automatically.
Claude can analyze your app's performance metrics, identify the slowest endpoints, find N+1 queries, and suggest concrete optimizations based on your actual data.
user: Why is /api/orders so slow?
claude: Analyzing the last 1000 requests to /api/orders...
Found the issue: An N+1 query is loading order items individually.
Average: 847ms (423ms in DB)
This should reduce response time to ~120ms.
Connect Pulse to Claude Code
npm install -g @brainzlab/pulse-mcp
// ~/.claude/config.json { "mcpServers": { "pulse": { "command": "pulse-mcp", "args": ["--api-key", "YOUR_API_KEY"] } } }
Open Claude Code and ask about performance:
"What are the slowest endpoints and how can I speed them up?"
Natural language performance optimization
What's slowing down my app the most?
Find top performance bottlenecks
Are there any N+1 queries I should fix?
Automatic query analysis
Compare performance before and after the last deploy
Deployment impact analysis
Which background jobs are taking too long?
Job performance monitoring
Show me the request trace for this slow endpoint
Detailed request breakdown
What database queries are running on this endpoint?
SQL analysis and optimization
Tools available to Claude
pulse_slow_endpoints
Find slowest endpoints by percentile
params: percentile, time_range, limit
pulse_get_trace
Get detailed request trace
params: trace_id
pulse_n1_queries
Detect N+1 query patterns
params: endpoint, time_range
pulse_compare
Compare performance between time ranges
params: endpoint, before, after
pulse_db_analysis
Analyze database query patterns
params: endpoint, time_range
pulse_throughput
Get throughput and latency stats
params: endpoint, time_range
Get started with Pulse and Claude Code today.