logo

lmnr-ai/lmnr

Laminar - open-source observability platform purpose-built for AI agents. YC S24.

View on GitHub ↗Jump to charts ↓Open shareable report →

Data as of . Signed-in members get hourly updates — create a free account.

Summary Information

Updated 34 minutes ago
Type:Server / PlatformCategory(s):LLM Evals, Guardrails & ObservabilityAgent Frameworks & SDKsAI Agents & LLM Apps
Added to GitGenius on September 21st, 2026
Created on August 29th, 2024
Open Issues & Pull Requests: 142 (+0)
GitHub issues: Enabled
Number of forks: 247
Total Stargazers: 3,364 (+0)
Total Subscribers: 9 (+0)

Repository Insights (GitGenius)

Median issue/PR response: 4.2 hours
Mean response time: 4.9 days
90th percentile: 4.9 days
Tracked items: 147

Most active contributors

Sign in to see contributor activity.
Sign in

Related repositories by overlapping contributors

Maintainer activity

5 people did triage or write work on this repository in the last 12 months.

Counts unlabeled, assigned, unassigned, milestoned, demilestoned, locked, unlocked over the last 12 months. These are issue and pull request events that require triage or write permission. Commits and code review are not counted. labeled and renamed are excluded because GitHub issue forms record the issue author as the actor. Figures from October 7, 2026. This count is not comparable across projects: each project's automation decides which of these events a person emits.

How this project is maintained

Roughly one issue in two opened in the past year never receives a reply. 94% of open issues come from outside the core team, so the backlog reflects real-world use rather than internal planning. Work labelled "question" is answered fastest, typically in about 5 hours, while "enhancement" waits about 10 days. 15% of tracked open issues have had no activity in three months. Only 35% of issues opened in the past year have been closed.

Charts & Analytics

Fetching additional details & charts...

Issue Activity (beta)

Open issues: 64
New in 7 days: 13
Closed in 7 days: 0
Avg open age: 155 days
Stale 30+ days: 27
Stale 90+ days: 19

Recent activity

Opened in 7 days: 13
Closed in 7 days: 0
Comments in 7 days: 2
Events in 7 days: 2

Top labels

  • frontend (25)
  • good first issue (23)
  • enhancement (11)
  • UX (10)
  • question (10)
  • UI (9)
  • bug (9)
  • help wanted (8)

Most active issues this week

Sign in to see which issues are moving.
Sign in
Explore full issue details

Detailed Description

Laminar is an open-source observability platform purpose-built for AI agents.

The platform addresses the challenge of monitoring and debugging AI agent behavior by providing integrated tracing, signal detection, and evaluation capabilities. Its tracing SDK uses OpenTelemetry and automatically instruments popular frameworks including Vercel AI SDK, Browser Use, Stagehand, LangChain, and major LLM providers with minimal setup. The signals feature allows developers to describe agent behaviors in plain English—such as "agent is stuck in a loop"—and receive Slack notifications when those conditions occur. For evaluation, the tool offers an unopinionated SDK and CLI that runs locally or in CI/CD pipelines, with a UI for visualizing and comparing results. The platform includes SQL query access to traces, spans, metrics, and events, enabling coding agents to investigate issues autonomously through MCP and CLI interfaces. Custom dashboards support visualization of traces, metrics, and events with custom SQL queries, while data annotation and dataset creation tools facilitate building evaluation datasets.

Developers should choose this tool if they need comprehensive visibility into agent execution and want to catch behavioral anomalies automatically. It suits teams building production AI agents who require both real-time monitoring and systematic evaluation workflows. The platform is available as a managed service or can be self-hosted via Docker Compose, with a full-featured production configuration available for on-premise deployments.

The project demonstrates active development with a focus on performance optimization. The codebase is written in Rust to achieve high throughput and implements 20x trace compression for efficient data ingestion and storage. The team has built a custom real-time engine for viewing traces as they happen and implemented ultra-fast full-text search over span data. The platform uses gRPC for trace data export, indicating attention to interoperability with standard observability infrastructure.