Our assessment
LangSmith is where LangChain makes its money, and it earns it: traces show every model call, tool call and intermediate step of an agent run, production dashboards and alerts sit on top, and any trace can be turned into a dataset for evaluation. For teams already on LangChain or LangGraph it is close to zero setup, and it is not locked to them: it lists integrations with the OpenAI and Anthropic SDKs, CrewAI, the Vercel AI SDK and Pydantic AI.
Plus is $39 per seat per month with 10k base traces included and usage-based pricing after that, so cost tracks both team size and traffic. It is also the route to LangSmith Deployment for hosting long-running agents. Self-hosting and hybrid setups are Enterprise-only, which matters if traces contain data that cannot leave your environment.
Pick it if you build on LangChain or LangGraph and want tracing and evals that just work. If you want open source and self-hosting without an enterprise contract, Langfuse is the closest alternative. For eval-first workflows, Braintrust is the main commercial competitor.
- Best for
- Tracing and evaluating LangChain/LangGraph agents, with a path to managed deployment.
- Watch out for
- Per-seat plus usage pricing; self-hosting requires Enterprise.
- Consider instead
- Langfuse (free, OSS) to self-host · Braintrust for evaluation-led workflows.
How it compares in Evals
History
Tracked since Oct 13, 2026.
| Snapshot | Rank | Adoption | Entry price |
|---|---|---|---|
| Nov 20, 2026 | #10 same | 86 | $39/seat |
| Nov 12, 2026 | #10 same | 86 | $39/seat |
| Oct 29, 2026 | #10 same | 86 | $39/seat |
| Oct 15, 2026 | #10 same | 86 | $39/seat |
| Oct 13, 2026 | #10 | 86 | $39/seat |
Plans
In US dollars, checked 15 Oct 2026.
| Developer | Free, 1 seat, 5k base traces / mo |
| Plus | $39 / seat / mo: 10k base traces, unlimited seats |
| Enterprise | Custom |
Key facts
| Category | Evals: observability and evaluation |
| Maker | LangChain |
| License | Proprietary |
| Self-host | Enterprise: self-hosted or hybrid |
| Free tier | Yes, Developer plan |
| Tracing | Traces, threads, dashboards, alerts |
| Evaluation | Datasets from traces, online evaluations |
| Feedback | Annotation queues and inline annotation |
| Deployment | LangSmith Deployment (Plus and up) |
| Framework support | LangChain/LangGraph plus OpenAI, Anthropic, CrewAI, Vercel AI SDK, Pydantic AI |
