LangSmith

What it is

LangChain's commercial observability platform.

LangSmith is the commercial platform from the LangChain team, and it offers the richest tracing for LangChain and LangGraph apps. It visualizes agent graphs natively, provides annotation queues for structured human review, and bundles evals and monitoring. If your stack is already LangChain, no other tool understands your runs as deeply.

The one job

Deep tracing and evals for LangChain apps.

It gives you polished, managed tracing and evaluation for LangChain and LangGraph applications, with the graph view and review workflows that native integration makes possible.

Reach for it when

• You build on LangChain or LangGraph and want the deepest native tracing.
• You want polished, managed tracing without running your own backend.
• You need annotation queues for structured human review of runs.
• You want an eval-driven workflow tied to your traces and datasets.

Skip it when

• You are not on LangChain: Langfuse or Phoenix are more neutral choices.
• You require self-host or open source: LangSmith is a commercial, managed product.
• You are on a tight budget at scale, where managed pricing adds up.

A minimal look

Tracing turns on via env vars.

// set once in the environment, no code change to your call:
// LANGSMITH_TRACING=true
// LANGSMITH_API_KEY=ls-...
import { wrapOpenAI } from "langsmith/wrappers";
import OpenAI from "openai";

const openai = wrapOpenAI(new OpenAI());  // every call is now traced
const res = await openai.chat.completions.create({
  model: "gpt-4o", messages,
});                                       // span appears in LangSmith

The principle it teaches

Where it fits your labs

This is a Week 5 tool. When your Lab 5 agent is built on LangGraph, LangSmith is the observability layer that renders the graph, captures each trace, and hosts your evals.