LangSmith Documentation
by LangChain
Trace, evaluate, and monitor LLM and agent apps in production — framework-agnostic.
Overview
LangSmith is LangChain's LLM observability and agent-engineering platform, and its documentation covers the full lifecycle of building reliable AI applications: capturing traces of every step in an LLM or agent run (retrieval, tool calls, intermediate transformations, and final output), running offline evaluations against datasets with built-in and custom evaluators including LLM-as-judge, collecting human and online feedback, versioning and testing prompts, and monitoring production performance through dashboards and alerting rules. It is explicitly framework-agnostic — the docs show integrations for the OpenAI SDK, Anthropic, CrewAI, Vercel AI SDK, Pydantic AI, and LlamaIndex — and supports cloud, hybrid, and self-hosted deployment. Python and TypeScript SDKs are documented with quickstarts that add tracing via environment variables or a decorator in minutes.
At a Glance
- Topic
- Frameworks
- Level
- All Levels
- Format
- Documentation
- Cost
- Freemium
- Duration
- Reference docs, self-paced
- Provider
- LangChain
- Hands-on
- Yes — code/exercises
- Certificate
- None
What You’ll Learn
- ✓Instrument LLM and agent apps to capture end-to-end traces of retrieval, tool calls, and model outputs
- ✓Build offline evaluation pipelines using datasets, built-in evaluators, and custom LLM-as-judge graders
- ✓Set up online evaluations, feedback collection, and automation rules for production monitoring
- ✓Version, test, and compare prompts systematically instead of eyeballing changes
- ✓Wire LangSmith into non-LangChain stacks (OpenAI SDK, Anthropic, LlamaIndex) via Python or TypeScript
Highlights
- •Framework-agnostic — works with any LLM provider or agent framework, not only LangChain
- •Covers observability, evals, prompt engineering, and deployment in one reference
- •Cloud, hybrid, and self-hosted options with a free tier to start
Who It’s For
Best For
- ✓AI engineers shipping LLM or agent apps to production who need tracing and evals
- ✓Teams standardizing on a single observability and evaluation workflow
- ✓Developers debugging multi-step agent and RAG pipelines
Prerequisites
- •Comfort building an LLM app in Python or TypeScript
- •Basic understanding of LLM API calls, tools, and agents
FAQ
What is LangSmith Documentation?
The official LangSmith documentation, a guide for AI engineers who need to debug, evaluate, and monitor LLM and agent applications in production. LangSmith is framework-agnostic observability, tracing, evals, and prompt engineering that works with any LLM stack, not just LangChain.
Is LangSmith Documentation free?
LangSmith Documentation offers free content, with paid options for certificates or premium features.
What level is LangSmith Documentation for?
LangSmith Documentation is aimed at a all levels audience. Recommended background: Comfort building an LLM app in Python or TypeScript, Basic understanding of LLM API calls, tools, and agents.
How long does LangSmith Documentation take?
Expect roughly Reference docs, self-paced. Most learners work through it at their own pace.
What will I learn from LangSmith Documentation?
You'll learn: Instrument LLM and agent apps to capture end-to-end traces of retrieval, tool calls, and model outputs; Build offline evaluation pipelines using datasets, built-in evaluators, and custom LLM-as-judge graders; Set up online evaluations, feedback collection, and automation rules for production monitoring; Version, test, and compare prompts systematically instead of eyeballing changes; Wire LangSmith into non-LangChain stacks (OpenAI SDK, Anthropic, LlamaIndex) via Python or TypeScript.