What is Laminar?
Open-source tracing and analysis for agents.
Who is Laminar best for?
Laminar is best for developers building AI agents who want every model and tool call traced and turned into evals.
Debugging agent runs
Each run is captured as a readable transcript of model calls, tool calls, sub-agents, tokens and cost.
Finding failures you did not expect
Signals read every run and group similar failures into clusters, so a new bug report can be matched to earlier ones.
Teams that want to self-host
The code is open source under Apache 2.0 and runs with Docker Compose, and Enterprise adds an on-premise option.
Browser agents
It records the agent's browser session next to the trace and works with Browser Use, Stagehand and Playwright.
How much does Laminar cost?
Laminar has a free plan with 1 GB of trace data, and its paid cloud plans cost $30 a month (Starter) or $150 a month (Pro), with extra data billed per GB.
- 1 GB of data and $5 in Signals credits, no overage
- 7-day retention
- 1 project, 1 seat
- 3 GB of data, then $2 per GB
- $15 in Signals credits, 30-day retention
- Unlimited projects and seats, email support
- 10 GB of data, then $1.50 per GB
- $50 in Signals credits, 6-month retention
- Unlimited projects and seats, Slack support
- Custom limits and rates
- On-premise deployment
- Dedicated support
Signals, the agent that reads traces to find failures, is billed per token once the included credits run out: $0.50 per million input and $3 per million output tokens on Starter, $0.40 and $2.50 on Pro. Data means the text and images Laminar stores across traces, evaluations and datasets. The Free plan stops at its caps instead of billing overage. Laminar is open source under Apache 2.0 and can be self-hosted. Full pricing on laminar.sh. We checked these prices on .
How do readers rate Laminar?
No one has rated Laminar yet. Be the first.
Sign in to rate. One rating per account.
Is Laminar open source?
Yes. Laminar's code is public under the Apache 2.0 licence, so you can run it on your own server.
On GitHub
3,485stars
+110 (+3.3%) in the last 30 days
- Last commit
- 12 Nov
- Repository age
- 2 years
- Latest release
- 10 Nov
- Licence
- Apache 2.0
- Self-hosted
- Yes
- Repository
- lmnr-ai/lmnr
What are the best Laminar alternatives?
Laminar is an open source alternative to Confident AI and Pydantic Logfire. AgentOps, Langfuse and SigNoz do the same job and are open source too.
Live traces turned into test cases for LLM apps.
One trace across model calls, agents, APIs and database queries.
Trace, debug and cost-track your AI agents in production.
Open-source tracing, prompt management and evals for LLM apps.
Observability platform with logs, metrics, traces and alerts in one place, built on OpenTelemetry.
Questions about Laminar
What is Laminar?
Open-source tracing and analysis for agents.
Who is Laminar for?
Laminar is best for developers building AI agents who want every model and tool call traced and turned into evals.
Is Laminar free?
Yes. Laminar has a free plan: 1 GB of data and $5 in Signals credits, no overage.
How much does Laminar cost?
Laminar's paid plans are Starter at $30 per month and Pro at $150 per month.
Is Laminar open source?
Yes. Laminar's code is public under the Apache 2.0 licence, so you can run it on your own server.
What are the best Laminar alternatives?
The closest alternatives on Recaply are Confident AI, Pydantic Logfire, AgentOps, Langfuse and SigNoz.
Where can I try Laminar?
On its official website, laminar.sh.
Can I self-host Laminar?
Yes. Laminar is open source under the Apache 2.0 license, and its GitHub repository runs locally with Docker Compose. For production, Laminar points to its managed cloud or its full Docker Compose setup, and self-hosted installs send anonymous usage data unless you switch it off.
What happens when I go over my Laminar plan's limits?
Paid Laminar plans keep running and bill extra data at $2 per GB on Starter or $1.50 per GB on Pro, plus Signals tokens at each plan's rates. The Free plan has no overage, so once its data cap or Signals credit runs out you need to upgrade.







