Warpmetrics Documentation

Warpmetrics is an observability platform for AI agents. Track every LLM call, measure costs, monitor latency, and classify outcomes across your agent workflows.

Quick Start

Get up and running with Warpmetrics in under 5 minutes.

Start the quickstart guide

Explore

Core Concepts

Understand runs, groups, calls, outcomes, and how they connect.

SDK Reference

Instrument your AI agents with @warpmetrics/warp in minutes.

MCP Integration

Connect AI assistants like Claude to query your metrics.

API Reference

Build custom integrations with the REST API.

How It Works

Wrap your LLM client

Use warp(client) to wrap your OpenAI or Anthropic client. All calls are automatically tracked.

Organize with runs and groups

Create runs for top-level workflows and groups for nested phases. Link calls to track hierarchy.

Record outcomes

Mark runs with outcomes like "completed" or "failed" to track success rates over time.

View in the dashboard

See costs, latency, success rates, and detailed call logs in real time.