AI Frameworks · lmnr-ai

lmnr

Laminar is an open-source observability platform designed to monitor and debug AI agents. It provides tracing, performance signals, evaluation tools, and dashboards to track agent behavior in real-time, with a Rust-powered backend optimized for high-volume trace ingestion.

Talk to DEV.co Discuss Implementation

Source: GitHub — github.com/lmnr-ai/lmnr

3.1k

GitHub stars

216

Forks

TypeScript

Primary language

Apache-2.0

License (OSI-approved)

Key facts

Objective fields from the source. Values we can't verify are shown as “Unknown” rather than guessed.

Field	Value
Repository	lmnr-ai/lmnr
Owner	lmnr-ai
Primary language	TypeScript
License	Apache-2.0 — OSI-approved
Stars	3.1k
Forks	216
Open issues	92
Latest release	v0.2.0 (2026-06-18)
Last updated	2026-07-08
Source	https://github.com/lmnr-ai/lmnr

What lmnr is

Built in TypeScript (frontend) and Rust (backend), Laminar offers OpenTelemetry-native tracing with automatic instrumentation for major LLM frameworks (LangChain, Vercel AI SDK, OpenAI, Anthropic), gRPC export, SQL query capabilities via MCP/CLI, and custom trace compression. Deployable as self-hosted Docker Compose or managed SaaS.

Quickstart

Get the lmnr source

Clone the repository and explore it locally.

terminalbash

1git clone https://github.com/lmnr-ai/lmnr.git2cd lmnr3# follow the project's README for install & configuration

Need it deployed, integrated, or customized instead? DEV.co ships production installs.

Best use cases

Debugging Multi-Step Agent Workflows

Trace end-to-end execution of complex agents across multiple LLM calls, tool invocations, and decision points. Real-time span visualization and full-text search over trace data accelerate root-cause analysis.

Evaluating Agent Quality at Scale

Run unopinionated evals locally or in CI/CD pipelines with custom metrics and visualize results side-by-side. Use datasets and annotation UI to create evaluation ground truth for iterative agent improvement.

Monitoring Agent Behavior in Production

Set up natural-language signals (e.g., 'agent stuck in loop') to automatically detect anomalies and alert via Slack. Dashboards with custom SQL queries provide business and operational metrics at a glance.

Implementation considerations

Install SDKs (npm/@lmnr-ai/lmnr or pip/lmnr) and add one-line initialization; auto-instrumentation covers OpenAI, Anthropic, LangChain, Vercel AI SDK out of the box.
For production self-host, use docker-compose-full.yml; configure LLM provider (Gemini, OpenAI, AWS Bedrock) for AI features (chat-with-trace, SQL-with-AI).
Custom Postgres schema support available; telemetry is anonymized but can be disabled with LAMINAR_TELEMETRY_DISABLED=true.
Evals require local SDK setup; CLI and custom render UI for data annotation indicate extensibility, but scope of custom evals logic is not clearly documented.
OpenTelemetry-native design allows future integration with broader observability ecosystems, though current integrations focus on LLM/agent frameworks.

When to avoid it — and what to weigh

Need Mature, Battle-Tested Observability — Project is ~10 months old (created Aug 2024). v0.2.0 release suggests pre-1.0 stability. Evaluate risk tolerance for production use of emerging platforms.
Require Multi-Tenancy and Enterprise Security Hardening — No explicit documentation provided on RBAC, audit logging, data isolation, or SOC 2 compliance. Requires detailed security review before enterprise deployment.
Already Heavily Invested in Competitor Stack — If your team standardizes on Datadog, New Relic, or similar, integration effort and training overhead may outweigh AI-specific features.
Cannot Self-Host or Prefer Vendor Lock-In — Managed platform at laminar.sh is primary path; self-hosted Docker Compose is lightweight and suitable for development/testing, not necessarily production scale.

License & commercial use

Apache License 2.0 (Apache-2.0). A permissive OSI-approved open-source license allowing commercial use, modification, and distribution, provided the license and any significant changes are documented.

Apache-2.0 is a permissive license generally compatible with commercial use. However, this does not extend to any managed services, SaaS terms, or proprietary components offered by the vendor (laminar.sh). Review laminar.sh terms of service separately for commercial use of the hosted platform. Self-hosted deployments can be used commercially under the Apache-2.0 terms.

DEV.co evaluation signals

Editorial assessment — not user reviews. Directional, with an explicit confidence level.

Signal	Assessment
Maintenance	Active
Documentation	Strong
License clarity	Clear
Deployment complexity	Low
DEV.co fit	Strong
Assessment confidence	High

Security considerations

No explicit security posture documentation provided. Self-hosted deployments require securing Postgres, gRPC endpoints, and LLM API keys in environment. Anonymous telemetry is default (opt-out available). Data sanitization policies, encryption at rest/in-transit, RBAC, and audit logging not documented—requires vendor review for sensitive agent workflows. Managed platform security model unknown.

Alternatives to consider

Langsmith (LangChain)

Purpose-built LLM/agent tracing and evaluation; mature, but tightly coupled to LangChain ecosystem. Broader framework support in Laminar via OpenTelemetry.

Arize

Established ML observability platform with evals and monitoring; more mature, but less AI-agent-specific and heavier operational footprint than Laminar.

Datadog APM + custom dashboards

Enterprise-grade observability with agent instrumentation possible via OpenTelemetry; more mature and battle-tested, but less AI-native and higher cost.

Software development agency

Build on lmnr with DEV.co software developers

Start with a free project at laminar.sh, or deploy self-hosted in minutes. Install the SDK and trace your first agent run in under 5 minutes.

Talk to DEV.co

Related open-source tools

Surfaced by semantic similarity across the DEV.co open-source index.

aichat openlit paddler

Related on DEV.co

Explore the category and the services that help you build with it.

AI Frameworks AI Development DevOps & Infrastructure Talk to DEV.co

lmnr FAQ

Does Laminar support my LLM framework?

Automatic instrumentation works for OpenAI, Anthropic, Gemini, LangChain, Vercel AI SDK, Browser Use, and Stagehand out of the box. Custom frameworks can use the OpenTelemetry SDK and observe() wrapper. See integrations at laminar.sh/docs/tracing/integrations.

Can I self-host in production?

Yes, via docker-compose-full.yml. Requires Postgres, LLM provider config, and proper infrastructure (scaling, backups, TLS). Managed platform at laminar.sh recommended for production unless you have DevOps capacity.

What are 'Signals'?

Natural-language behavioral patterns (e.g., 'agent stuck in loop') that Laminar monitors across all agent runs. When detected, automated Slack alerts notify your team. No need to define thresholds manually.

Is my data private if I self-host?

Self-hosted deployments store data in your own Postgres instance. Managed platform (laminar.sh) privacy depends on vendor terms (not provided in data). Anonymous telemetry is collected unless LAMINAR_TELEMETRY_DISABLED=true is set.

Software developers & web developers for hire

Adopting lmnr is usually one piece of a larger software development effort. As a software development agency, DEV.co provides software development services and web development expertise — pairing senior software developers and web developers with your team to design, build, and operate ai frameworks software in production.

Ready to instrument your AI agents?

Start with a free project at laminar.sh, or deploy self-hosted in minutes. Install the SDK and trace your first agent run in under 5 minutes.

Talk to DEV.co Schedule a Consultation