Complete AI agent control of your observability stack. From setup to investigation to remediation—your AI assistant can handle it all. Works with Cursor, Claude Desktop, and any MCP-compatible agent.
For Cursor, Claude Desktop, and any MCP-compatible agent. Full access to all observability tools.
Learn more →Inline error annotations, service health sidebar, and trace viewer directly in your editor.
Learn more →PR annotations with production errors, deployment tracking, and automated health checks.
Learn more →JSON output for all commands, perfect for piping to AI agents or scripts.
Learn more →Direct API endpoints optimized for AI agents with structured, context-rich responses.
Learn more →Pre-built investigation guides that AI agents can follow step-by-step.
Learn more →The Model Context Protocol (MCP) server gives AI agents like Claude and Cursor direct access to your observability data. Ask questions in natural language, and your AI can query logs, search traces, and investigate issues.
// .cursor/mcp.json
{
"mcpServers": {
"qorrelate": {
"command": "npx",
"args": ["qorrelate-mcp-server"],
"env": {
"QORRELATE_API_KEY": "your-api-key",
"QORRELATE_ENDPOINT": "https://qorrelate.io"
}
}
}
}"Show me all errors from the payment service in the last hour"
→ Uses query_logs with severity=ERROR and service=payment
"What's causing the latency spike in checkout?"
→ Uses search_traces with min_duration and traces_analyze_latency
"Are there any critical alerts I should know about?"
→ Uses list_alerts with severity=critical
"What's my current usage and billing this month?"
→ Uses get_billing and get_usage to show cost breakdown
"Drop all DEBUG logs from production to reduce costs"
→ Uses create_drop_filter with filter_type=logs and query="level:DEBUG"
"Find unused metrics I can drop to save money"
→ Uses get_unused_metrics to identify metrics not queried in 7 days
See production errors directly in your code. The VS Code extension matches stack traces from production to your local files and shows inline annotations.
ext install qorrelate.qorrelateThe Qorrelate CLI includes agent-optimized commands with JSON output, perfect for piping to AI agents or automation scripts.
# Get service health as JSON
qorrelate agent health --format json
# Search logs with JSON output
qorrelate agent logs --severity ERROR --format json
# Natural language query
qorrelate agent query "slow traces in checkout" --format json
# Get investigation runbook
qorrelate agent runbook high_error_rate --format json
Install the CLI: curl -sL https://install.qorrelate.io | sh
Direct API endpoints optimized for AI agents. Responses are structured for easy parsing and include context, recommendations, and follow-up suggestions.
GET /v1/agent/services/summary
Get health summary for all services with recommendations
GET /v1/agent/investigate
Get investigation context for a trace ID or error message
GET /v1/agent/query
Natural language query processing
GET /v1/agent/runbook/{name}
Get step-by-step investigation runbook
Pre-built investigation guides that AI agents can follow step-by-step. Each runbook includes the tools to use, parameters to pass, and common causes to check.
Steps to diagnose elevated error rates
Debug latency issues and bottlenecks
Investigate service unavailability
Post-deployment health checks
Get started with Qorrelate's AI integrations in minutes.