Open-source observability platform for AI agents with trace capture, step-restart debugging, browser session recording, and natural language pattern detection. Self-host free or use managed cloud from $30/month.
Open-source monitoring for AI agents. Trace every step, debug failures by restarting from any point, record browser sessions, and catch problems with natural language pattern matching.
Laminar (LMNR) is an open-source AI agent observability platform for tracing, debugging, evaluating, and improving long-running agents; teams can self-host for free, use a Free cloud tier, or choose managed cloud plans starting at $30/month, with Pro listed at $150/month and Enterprise priced custom.
It is built for developers working on complex agent systems where failures can happen late in a run and simple request-response logging is not enough. The platform's core positioning is agent observability rather than generic application monitoring: it focuses on showing how an agent moved through its steps, what tools it called, what browser actions it took, what model responses shaped its decisions, and where the workflow began to drift from the expected path.
The most distinctive workflow is Laminar's agent debugger. The product describes a rerun-at-step-N experience that preserves previous context, which is useful when a research, coding, browser, or operations agent fails after many minutes of work. Instead of repeatedly replaying a full run from the beginning, a developer can inspect the trace around the failing decision point and iterate from that step. This makes Laminar especially relevant for teams building long-running agents, browser agents, multi-tool assistants, and evaluation-heavy development loops where the expensive part is not just logging a request but understanding the sequence of decisions that led to a bad outcome.
Laminar also includes browser session recording synced to traces. That matters for web automation agents because a trace alone may show the tool call or action, while the recording can reveal what was actually visible in the browser when the agent clicked, navigated, waited, or extracted information. The listed integrations include Browser Use, Stagehand, Playwright, Kernel, and Browserbase, which gives the product a clear fit for teams testing and operating agents that interact with live web pages. For non-browser workflows, Laminar still provides trace capture, cost and latency visibility, evaluation support, dataset-related workflows, SQL querying, and dashboards for reviewing operational behavior.
Signals are Laminar's natural-language pattern detection feature. Instead of relying only on manually written log queries, teams can describe a failure pattern or operational event, define the desired output structure, and use Laminar to extract matching events from traces. This is useful for recurring issues such as retry loops, repeated failed actions, missing final outputs, tool errors, or user-visible frustration. Signals should still be treated as an analysis aid rather than a replacement for careful operational design: teams need to define useful categories, review the extracted examples, and connect the results to debugging or product-quality workflows.
Pricing is structured around a free cloud tier, paid managed cloud tiers, and custom Enterprise terms. The public pricing page lists Free at $0/month with 1 GB data, 1,000 Signals steps, 15-day retention, 1 project, 1 seat, and community support. Hobby is listed at $30/month with 3 GB data included, $2/GB overage, 5,000 Signals steps included, $0.0075 per additional Signals step, 30-day retention, unlimited projects, unlimited seats, and email support. Pro is listed at $150/month with 10 GB data included, $1.50/GB overage, 50,000 Signals steps included, $0.005 per additional Signals step, 90-day retention, unlimited projects, unlimited seats, and Slack support. Enterprise is custom priced with custom limits, on-premise deployment, unlimited projects, unlimited seats, and dedicated support. Teams should confirm current plan terms before purchase, but the published tier structure makes the entry path clear while reserving deployment and large-scale requirements for Enterprise discussions.
Was this helpful?
Laminar is a strong debugging tool for complex AI agents. The step-restart debugger and browser session recordings address problems that general-purpose observability tools often do not cover. Self-host for free or use managed cloud starting at $30/month. It is a young platform with a growing ecosystem and is best suited for teams building agents that chain multiple tools, browser actions, and evaluation workflows.
Restart a failed agent run from any step with full context. LLM calls replay from cached responses, external state such as browser sessions or sandboxes can be restored where supported. No full rerun needed.
Use Case:
An agent fails 40 minutes into a multi-step research task. Instead of rerunning the entire thing, restart from the exact decision point that went wrong and iterate on the fix.
Instrument supported frameworks and SDKs to capture inputs, outputs, token counts, latency, and cost for agent and LLM calls.
Use Case:
Get production visibility into an agent's behavior and cost by adding tracing instrumentation to the workflow.
Captures screen recordings from browser agents and syncs them with trace timelines. Integrates with Browser Use, Stagehand, Playwright, Kernel, and Browserbase.
Use Case:
Debug why a browser automation agent clicked the wrong button by watching the recording alongside the agent's decision trace.
Describe a failure pattern in plain English and Laminar finds matching instances across production traces. Runs continuously against new data where configured.
Use Case:
Find every instance where an agent entered a retry loop or a user expressed frustration, without writing custom log queries.
Run LLM-as-judge, deterministic, or custom evaluation functions against traces or curated datasets. Results can be tracked over time for regression detection.
Use Case:
Nightly evaluations against a golden dataset catch quality drops in a customer support agent before users report problems.
Query platform data with SQL. Feed evaluation inputs from SQL queries and pull data into external applications via SQL API where supported.
Use Case:
Build custom analytics correlating token usage with user satisfaction across different agent versions and prompt configurations.
$0/month
$30/month
$150/month
Custom
Ready to get started with Laminar (LMNR)?
View Pricing Options →We believe in transparent reviews. Here's what Laminar (LMNR) doesn't handle well:
Weekly insights on the latest AI tools, features, and trends delivered to your inbox.
LLM Observability
Langfuse is an open-source LLM observability and engineering platform providing tracing, prompt management, evaluations, and dataset management for production AI applications.
AI Observability
LangSmith is LangChain's commercial observability, evaluation and prompt management platform for LLM apps and agents in production.
LLM Observability
Open-source LLM observability and AI gateway — logs every prompt, response, cost, and latency across 20+ providers with a one-line proxy or async SDK, plus caching, retries, and prompt experiments.
AI Observability
Phoenix is Arize's open-source LLM observability project, and it has quietly become the default way tens of thousands of teams see what their agents are actually doing in production. The pitch is simple: `pip install arize-phoenix`, instrument with OpenInference (or any OpenTelemetry-compatible library), and every LLM call, tool invocation, retrieval, and embedding shows up as a spanned timeline you can filter, search, and replay. No vendor account required, no proprietary SDK lock-in. The Open
LLM Evaluation & Observability
LLM reliability platform that turns evals and monitors into a continuous feedback loop — recently announced to be joining ServiceNow.
No reviews yet. Be the first to share your experience!
Get started with Laminar (LMNR) and see if it's the right fit for your needs.
Get Started →Take our 60-second quiz to get personalized tool recommendations
Find Your Perfect AI Stack →Explore 20 ready-to-deploy AI agent templates for sales, support, dev, research, and operations.
Browse Agent Templates →