laminar

Laminar

English | 中文

Laminar is an open-source observability platform purpose-built for AI agents. It provides tracing, evaluations, AI monitoring, SQL access to data, dashboards, and data annotation capabilities.

Features

• Tracing: OpenTelemetry-native powerful tracing SDK - 1 line of code to automatically trace Vercel AI SDK, Browser Use, Stagehand, LangChain, OpenAI, Anthropic, Gemini, and more.
• Evals: Unopinionated, extensible SDK and CLI for running evals locally or in CI/CD pipeline.
• AI Monitoring: Define events with natural language descriptions to track issues and custom behavior.
• SQL Access: Query traces, metrics, and events with a built-in SQL editor.
• Dashboards: Powerful dashboard builder for traces, metrics, and events.
• Data Annotation: Custom data rendering UI for fast data annotation and dataset creation.
• High Performance: Written in Rust with custom realtime engine and ultra-fast full-text search.

Services

Service	Description	Port
frontend	Laminar web UI and Next.js application	5667
app-server	Rust backend server (HTTP API, gRPC, Real-time)	8000-8002
postgres	PostgreSQL database for persistent storage	-
clickhouse	ClickHouse for high-performance analytics	-
quickwit	Quickwit for full-text search	7280
query-engine	Query engine for data processing	8903

Ports

Port	Protocol	Description
5667	HTTP	Laminar Web UI
8000	HTTP	App Server HTTP API
8001	gRPC	App Server gRPC
8002	HTTP	App Server Real-time API
7280	HTTP	Quickwit REST API and UI
8903	HTTP	Query Engine API

Environment Variables

Variable	Description	Default
LAMINAR_POSTGRES_VERSION	PostgreSQL image version	16-alpine
LAMINAR_POSTGRES_USER	PostgreSQL username	postgres
LAMINAR_POSTGRES_PASSWORD	PostgreSQL password	change_me_in_production
LAMINAR_CLICKHOUSE_VERSION	ClickHouse image version	latest
LAMINAR_CLICKHOUSE_USER	ClickHouse username	clickhouse
LAMINAR_CLICKHOUSE_PASSWORD	ClickHouse password	change_me_in_production
LAMINAR_SHARED_SECRET_TOKEN	Shared secret for inter-service auth	shared_secret_change_in_production
LAMINAR_AEAD_SECRET_KEY	AEAD encryption key (64 hex chars)	-
LAMINAR_OPENAI_API_KEY	OpenAI API key (optional)	-
LAMINAR_GOOGLE_GENERATIVE_AI_API_KEY	Google AI key for Signals feature	-
LAMINAR_FRONTEND_PORT_OVERRIDE	Host port for Web UI	5667
LAMINAR_APP_SERVER_PORT_OVERRIDE	Host port for HTTP API	8000
LAMINAR_APP_SERVER_GRPC_PORT_OVERRIDE	Host port for gRPC	8001
LAMINAR_APP_SERVER_RT_PORT_OVERRIDE	Host port for Real-time API	8002

Volumes

• laminar_postgres_data: PostgreSQL data
• laminar_clickhouse_data: ClickHouse data
• laminar_clickhouse_logs: ClickHouse logs
• laminar_quickwit_data: Quickwit data

Quick Start

1. Copy the example environment file:

   cp .env.example .env

2. Edit .env and set required secrets:

   LAMINAR_POSTGRES_PASSWORD=your_secure_password
   LAMINAR_CLICKHOUSE_PASSWORD=your_secure_password
   LAMINAR_SHARED_SECRET_TOKEN=your_shared_secret
   LAMINAR_AEAD_SECRET_KEY=your_64_character_hex_key

   To generate a secure AEAD key:

   openssl rand -hex 32

3. Start the services:

   docker compose up -d

4. Access the UI at http://localhost:5667

Enabling AI Monitoring (Signals)

To enable the Signals feature, set your Google Generative AI API key:

LAMINAR_GOOGLE_GENERATIVE_AI_API_KEY=your_key_here

SDK Configuration

When using Laminar SDK with self-hosted instance:

Python:

from lmnr import Laminar
Laminar.initialize(
    project_api_key="<LMNR_PROJECT_API_KEY>",
    base_url="http://localhost:8000"
)

TypeScript:

import { Laminar } from '@lmnr-ai/lmnr';
Laminar.initialize({
    projectApiKey: process.env.LMNR_PROJECT_API_KEY,
    baseUrl: "http://localhost:8000"
});

Documentation

For full documentation, visit: https://laminar.sh/docs

Resource Requirements

Minimum recommended resources:

Service	CPU	Memory
frontend	0.25	256M
app-server	0.5	512M
postgres	0.25	256M
clickhouse	0.5	512M
quickwit	0.25	256M
query-engine	0.25	128M

Total minimum: ~2 CPU cores, ~2GB RAM

For production use, increase ClickHouse and PostgreSQL resources based on data volume.