Laminar
An open-source AI agent observability platform with Docker self-hosting, Apache 2.0 licensing, automated failure detection, and regression evals.
Laminar is an open-source platform for developers who need to see why an AI agent failed and check whether a fix worked. You can self-host it with Docker or on Kubernetes, including AWS and GCP, or use its managed cloud service. It uses the Apache 2.0 license.
Its traces capture LLM calls, tool use, sub-agents, token usage, and costs in a readable transcript. Browser agents can also attach session recordings. Signals analyze runs for unexpected failures and group similar cases, so teams can see recurring problems and check whether a reported issue has happened before.
The debugging and evaluation tools share that trace data:
- Failure clusters can become eval datasets for regression checks. Evals run locally or in CI, with a UI for comparing results.
- MCP and CLI access let coding assistants query traces and metrics with SQL, investigate failures, and verify fixes.
- Slack alerts report failures and clusters, and you can ask Laminar about traces there.
Laminar uses OpenTelemetry for tracing and integrates with Claude Agent SDK, OpenAI Agents SDK, Vercel AI SDK, LangChain, and Browser Use. It also has full-text search, SQL dashboards, and annotation tools for evaluation and fine-tuning datasets.
Self-hosting puts the stack on your infrastructure; the managed service runs on Laminar's. AI-assisted trace chat, SQL generation, and server workers require an LLM provider. Self-hosted deployments collect anonymized usage telemetry, which you can disable.
