Laminar

Laminar is an open source observability platform for AI agents. It records what an agent does during each run and presents it so developers can find and fix failures, rather than reading raw span trees. The platform automatically captures LLM calls, tool calls, sub agents, costs and tokens, and shows them in a transcript view that reads like a conversation, with a tree view, a timeline and custom render templates as alternatives. A feature called Signals analyzes every agent run to surface failure modes the team did not define in advance, and groups similar failures into clusters so developers can see whether a newly reported issue has happened before. Laminar also supports evaluations, and it can be self hosted. It is aimed at teams building and operating AI agents who need tracing, debugging and failure analysis in one place. Pricing is published in four plans. Free costs $0 per month and includes 1 GB of data, 7 day retention, 1 project, 1 seat and community support. Starter costs $30 per month and includes 3 GB, then $2 per GB, with 30 day retention, unlimited projects and seats, and email support. Pro costs $150 per month and includes 10 GB, then $1.50 per GB, with 6 month retention and Slack support. Enterprise is custom priced and adds on premise deployment and dedicated support. Each plan includes a one time $5 Signals credit, and Signals analysis beyond that credit is billed by tokens used on paid plans.