LangSmith
LLM observability + evals from the LangChain team — production tracing for AI apps.
Visit LangSmith Free 5k traces/mo, Plus $39/mo
What is LangSmith?
LangSmith is the production observability platform from the LangChain team. Trace every LLM call, debug failed agent runs, build eval datasets, A/B test prompts. Works with any LLM framework, not just LangChain.
Key features
- Full LLM call tracing
- Eval datasets + scoring
- Prompt playground + versioning
- A/B testing flows
- Framework-agnostic SDK
- Self-hostable enterprise option
Pros
- Industry-standard tracing for LangChain users
- Strong eval framework
- Free tier good for prototyping
Cons
- Tied to LangChain ergonomics (still works without)
- Self-hosted is enterprise-only
- Some features paywalled
Best for
LangChain developersTeams debugging agent workflowsAI engineers building eval pipelinesAnyone tracing LLM production traffic
Alternatives to LangSmith
Coding
Vellum
LLM dev platform — prompt versioning, evals, monitoring in one tool.
LLM dev platform
PaidFrom $500/mo
Released July 2023
Coding
Helicone
Open-source LLM observability — one-line proxy, full request logs, cost tracking.
LLM observability
FreemiumFree 100k requests/mo, Pro $25/mo
Released February 2023