With SigNoz, you can debug production without leaving your editor and explore your telemetry by asking questions in plain English. Use Noz, the AI teammate built into SigNoz, or connect your own coding agents to your data. Learn more about Agent Native Observability.
Noz: SigNoz's AI Teammate
Noz is SigNoz's AI teammate, built into the product. Open it as a sidepane while you work, or full screen to dig in. Ask about your logs, traces, and metrics in plain English, and Noz investigates across your telemetry and answers grounded in your real data. It can also create dashboards, alerts, and views when you ask.
- Talk to your observability stack in natural language, no query language required
- Investigate incidents, explain dashboards, and explore data without writing queries
- Agentic: reads your data, runs queries across signals, proposes fixes, and creates dashboards, alerts, and views
- Available to SigNoz Cloud users
Integration Options
Connect your own coding agents to SigNoz and go from alert to root cause to fix without leaving your dev environment.
MCP Server
The SigNoz MCP Server implements the Model Context Protocol (MCP) and connects SigNoz to your coding agents. Plug into Claude Code or Cursor in minutes and get full observability context in every session.
- Query metrics, logs, and traces in natural language
- Pull traces, logs, metrics and infrastructure data in one query, with your codebase already loaded
- Works with Claude, Cursor, Gemini, Codex, and other MCP compatible agents
Get started with the MCP Server →
Agent Skills
SigNoz Agent Skills are an open-format library of SKILL.md files that teach coding agents how to work with SigNoz. Your agent picks the right skill whenever a task calls for it - no prompting required.
- Look up docs, generate queries, build and explain dashboards, create and triage alerts, and manage saved views
- Install everything as the SigNoz plugin, or add individual skills via skills.sh
- Works with Claude Code, Codex, Cursor, Gemini, and other
SKILL.mdcompatible agents
Get started with Agent Skills →
AI Use Cases
Practical workflows for Noz and the MCP Server:
- Ask Noz what changed after a deploy when errors or latency jump.
- Have Noz read a noisy alert's history and recommend a saner threshold.
- Search and analyze logs by asking questions in plain English.
- Ask "why is this slow?" and get a span breakdown with the bottleneck identified.
