# AgentOps

An observability and devtool platform for AI agents, giving Python developers session replay, LLM cost tracking, and integrations with agent frameworks such as CrewAI, AutoGen and LangGraph.

AgentOps helps developers build, evaluate and monitor AI agents from prototype to production. Initializing its client and wrapping a program records session replay analytics with step-by-step execution graphs, so an agent's calls can be inspected after the fact.

It tracks spend across LLM providers, integrates natively with CrewAI, AG2 (AutoGen), Agno and LangGraph, and can be self-hosted as a full dashboard-plus-API-backend app instead of the hosted service. Its Python SDK adds decorators for session, agent, operation, task and workflow spans, each supporting input/output recording, exception handling, async functions and generators.

It installs with `pip install agentops`; the app is open source under the MIT license.

Website: https://agentops.ai