# AgentOps — Observability for AI Agents

> Python SDK for AI agent monitoring. LLM cost tracking, session replay, benchmarking, and error analysis. Integrates with CrewAI, LangChain, AutoGen, and more. 5.4K+ stars.

## Install

Save as a script file and run:

## Quick Use

```bash
pip install agentops
```

```python
import agentops
agentops.init(api_key="YOUR_KEY")

# Your existing agent code — AgentOps auto-instruments LLM calls
from openai import OpenAI
client = OpenAI()
response = client.chat.completions.create(
    model="gpt-4o",
    messages=[{"role": "user", "content": "Hello"}],
)

agentops.end_session("Success")
```

---

## Intro

AgentOps is a Python SDK for monitoring and observing AI agents in production. It auto-instruments LLM calls to track costs, latency, token usage, errors, and session replays. Integrates with CrewAI, LangChain, AutoGen, OpenAI Agents SDK, Agno, and CamelAI with one line of code. 5,400+ GitHub stars, MIT licensed.

**Best for**: Teams running AI agents in production who need cost control and debugging
**Works with**: OpenAI, Anthropic, CrewAI, LangChain, AutoGen, Agno, CamelAI

---

## Key Features

### Auto-Instrumentation
Add `agentops.init()` and all LLM calls are automatically tracked — no code changes needed.

### Cost Tracking
Real-time cost per session, per agent, per model. Set budget alerts.

### Session Replay
Replay entire agent sessions step-by-step: prompts, responses, tool calls, errors.

### Error Analysis
Automatic detection of hallucinations, infinite loops, and failed tool calls.

### Benchmarking
Compare agent performance across models, prompts, and configurations.

### Dashboard
Web dashboard for team-wide monitoring, analytics, and alerting.

---

### FAQ

**Q: What is AgentOps?**
A: A Python SDK for AI agent observability — cost tracking, session replay, benchmarking, and error analysis. Auto-instruments LLM calls. 5.4K+ stars.

**Q: Does AgentOps add latency?**
A: Minimal — instrumentation is async and non-blocking. Typically <5ms overhead per LLM call.

---

## Source & Thanks

> Created by [AgentOps AI](https://github.com/AgentOps-AI). Licensed under MIT.
> [AgentOps-AI/agentops](https://github.com/AgentOps-AI/agentops) — 5,400+ GitHub stars

---
Source: https://tokrepo.com/en/workflows/d570c84f-4e22-4723-806a-d23710686a5c
Author: Script Depot