LangSmith
LangChain's platform for tracing, evaluating, and monitoring LLM apps — framework-agnostic.
LangSmith is LangChain's hosted platform for tracing, evaluating, and monitoring LLM applications. It captures every step of a chain or agent run, lets you build datasets and run offline/online evals, and works whether or not you use LangChain.
LangSmith is LangChain's platform for the operational side of LLM apps: tracing every step of a run, evaluating against datasets, and monitoring quality, latency, and cost in production. Despite the name, it is framework-agnostic — you can instrument an app built with or without LangChain.
It is aimed at teams who want one place to see what their chains and agents actually did, turn real traffic into evaluation datasets, and catch regressions before and after deploy. Its tracing is especially useful for agents, where a single user request can fan out into many tool calls and model invocations that are hard to debug from logs alone.
Highlights
- Tracing — capture the full tree of LLM calls, tool calls, and intermediate steps for any run.
- Datasets & evals — build datasets from traces, run offline evals (including LLM-as-judge), and compare versions.
- Online evaluation & monitoring — score production traffic and track quality/latency/cost over time.
- Prompt management & playground — version prompts and iterate with a hosted playground.
- Framework-agnostic — SDKs and OpenTelemetry-style instrumentation for any stack.
In an AI-assisted workflow
Set a few environment variables and your runs start showing up as traces:
export LANGSMITH_TRACING=true
export LANGSMITH_API_KEY=...Then promote interesting traces into a dataset and run evaluations against it as you change prompts or models.
NOTE
Tracing is the foundation for everything else: you can't evaluate or debug what you can't see. Instrument first, then build datasets from real traffic.
Good to know
LangSmith is a commercial platform with a free tier and usage-based paid plans. It is hosted (a self-hostable enterprise option exists). For fully open-source alternatives, compare Langfuse and Arize Phoenix; see Best LLM & RAG Evaluation Tools in 2026 for the full landscape.
Frequently asked questions
- What is LangSmith?
- LangSmith is LangChain's platform for the operational side of LLM apps: tracing every step of a run, evaluating against datasets, and monitoring quality, latency, and cost in production. Despite the name, it is framework-agnostic — you can instrument an app built with or without LangChain.
- How much does LangSmith cost?
- LangSmith is a commercial platform with a free tier and usage-based paid plans. It is hosted, and a self-hostable enterprise option exists.
- How do I use LangSmith?
- Set a few environment variables — LANGSMITH_TRACING=true and LANGSMITH_API_KEY — and your runs start showing up as traces. From there, promote interesting traces into a dataset and run evaluations against it as you change prompts or models. Tracing is the foundation: instrument first, then build datasets from real traffic.
- LangSmith vs Langfuse?
- LangSmith is the commercial, hosted option from LangChain. For fully open-source alternatives, the usual comparisons are Langfuse and Arize Phoenix.
Related
- LangfuseAn open-source LLM engineering platform for tracing, evals, prompt management, and metrics.
- Arize PhoenixAn open-source LLM observability and evaluation tool built on OpenTelemetry, runnable anywhere.
- Best LLM & RAG Evaluation Tools in 2026: DeepEval vs RAGAS vs LangSmith vs Phoenix vs promptfooA decision guide to the LLM eval landscape — code-first frameworks vs. eval-and-observability platforms, open-source vs. hosted, and which fits your stack.
- LLM Observability EngineerUse this agent to make a production LLM app observable — tracing every step, scoring live traffic with online evals, and monitoring quality, cost, and latency — so you can debug agent runs and catch regressions in production. Examples — "add tracing to our RAG/agent so we can debug bad answers", "set up online evals and cost/latency dashboards", "production quality is slipping and we're flying blind".
- Langfuse vs LangSmith: LLM Observability Compared (2026)Langfuse vs LangSmith — open-source self-hostable observability vs LangChain's first-party platform. Tracing, evals, prompt management, and which to adopt.
- BraintrustAn end-to-end platform for evaluating, iterating on, and observing LLM apps, with a prompt playground.
- LangchainThe provider-agnostic agent framework, post-1.0: a standard create_agent loop on the LangGraph runtime, middleware hooks, and the largest integration ecosystem.