Engineers debugging agents get step-by-step replays and cost tracking free up to 5,000 events, and the open-source app can be self-hosted.

Kenning's take

What it does

AgentOps is an observability tool for developers building AI agents and LLM apps. After adding its SDK, it records each LLM call, tool use and multi-agent interaction so developers can replay runs step by step, track token use and cost across 400+ models, and keep a log of errors and prompt injection attempts. It integrates with frameworks such as OpenAI Agents SDK, CrewAI, AutoGen and LangGraph, and the app is open source and can be self-hosted.

• Session replay and time-travel debugging
• LLM cost and token tracking across 400+ models
• Integrations with CrewAI, AutoGen, LangGraph and OpenAI Agents SDK
• Python and TypeScript SDKs
• Open-source app with self-hosting option

Who it suits. Engineers building and running AI agents who need to see why an agent failed and what it is costing.

Key facts

Category

Developer tools

Made by

Agency (AgentOps)

How to get it

sign up free on the website and install the SDK (pip install agentops); Enterprise via demo

Works on

Web, Python, TypeScript

In the UK

Yes

Free plan

Basic: free up to 5,000 events

Pricing

Plan	Price	Billed
Basic	$0	month
Pro	$from 40	month
Enterprise	Custom pricing n/a	n/a

source

FAQ

Facts from the maker's own pages and the sources listed. Checked 2026-10-11.