Home / Agent Debugging & Replay
AgentOps
Session-replay style observability for AI agents with time-travel debugging, cost tracking and first-class integrations into agent frameworks like CrewAI and AutoGen.
| Category | Agent Debugging & Replay |
| OpenTelemetry-native | Yes — no proprietary-SDK lock-in |
| Open source | Yes |
| Self-hostable | No |
| Pricing model | freemium |
| Pricing notes | Free tier + paid plans; SDK is MIT OSS |
| Framework integrations | crewai, autogen, openai-sdk, langchain, camel |
| Funding / ownership | $2.6M pre-seed |
| GitHub stars | 5,629 |
| Maintenance | slow (2026-03-19) |
| License (GitHub) | MIT |
| Open issues | 163 |
| Maturity signal | 75/100 (Established) — computed from public GitHub signals, see methodology |
Pricing/feature source: https://www.agentops.ai
Maturity signal is computed from public GitHub data only. How it is calculated.
Frequently asked questions
Is AgentOps open source?
Yes. Licensed under MIT.
Can I self-host AgentOps?
No — AgentOps is cloud-only; there is no self-hosted deployment.
Is AgentOps OpenTelemetry-native?
Yes — AgentOps speaks OpenTelemetry (OTLP) natively, so there is no proprietary-SDK lock-in.
How much does AgentOps cost?
Freemium. Free tier + paid plans; SDK is MIT OSS