About AgentOps
AgentOps records what AI agents do, so developers can trace, debug and audit them. It works with many agent frameworks and LLM providers through one SDK.
What it does
AgentOps logs LLM calls, tool calls and multi-agent actions as one timeline. You can replay an agent run and step through it to find where things went wrong. It tracks token counts and cost across more than 400 LLMs, and it keeps a full trail of logs and errors for review.
Key features
- Timeline of LLM calls, tools and agent actions
- Replay and time travel debugging of agent runs
- Token and cost tracking across 400+ models
- Session replay for agent runs
- Session and event export on paid plans
Who it's for
Developers building AI agents with frameworks such as OpenAI, CrewAI or AutoGen. Teams running multi-agent systems in production will also find the audit trail useful.
Pricing and good to know
The SDK is MIT licensed and available on GitHub. The Basic plan is free and includes up to 5,000 events. The Pro plan starts at $40 per month and adds unlimited events and log retention. The Enterprise plan is priced on request and adds options such as custom SSO and self-hosting.
Highlights
- View LLM calls, tools and agent actions in one timeline
- Replay agent runs step by step
- Track token counts and cost per model
- Export sessions and events on paid plans
