Comparison
Langfuse vs LangSmith
Langfuse and LangSmith both trace what your agents do. Langfuse is the open-source observability platform you can self-host, built around traces, prompts, and evaluations. LangSmith is LangChain hosted platform for tracing, evaluation, and dataset-driven testing of agent apps.
Side by side
| Dimension | Langfuse | LangSmith |
|---|---|---|
| Pricing model | Freemium (hosted paid tiers) | Freemium (hosted paid tiers) |
| License / terms | MIT | Not open source |
| Deployment | cloud, self-hosted, docker, kubernetes | cloud, bring-your-own-cloud, self-hosted |
| MCP support | The Langfuse homepage states that SKILL.md, CLI, and MCP connect coding agents to Langfuse for work in the app or IDE. | Current MCP support details were not cleanly verified for Braintrust in this research pass. Check the official Braintrust documentation for the supported setup before wiring an MCP client to it. |
| Primary package | PyPI: langfuse 4.17.0 | PyPI: langsmith 0.14.4 |
| Authentication | API keys with public and secret project keys via environment variables | API key via LANGSMITH_API_KEY environment variable, with optional workspace ID for org-scoped keys |
| Vendor | Langfuse | LangChain |
AgentsUse has not run a head-to-head benchmark, so this page claims no winner. It maps the differences your job has to price.
Every cell comes from the verified profiles, which trace to the official repository, package page, and documentation checked on the dates shown there. Pricing cells name the model, not a price: vendor prices change, so check each profile for the sourced pricing summary and official pricing link.
Differences that decide the choice
- Openness: Langfuse core is open source and self-hostable. LangSmith is a hosted LangChain product.
- Framework gravity: LangSmith is the native observability layer for LangChain and LangGraph apps. Langfuse is framework-agnostic with SDKs and an OpenAI-compatible API surface.
- Focus: both cover traces and evaluations; Langfuse leans into prompt management and cost views, LangSmith into datasets and offline evaluation tied to the LangChain stack.
Choose Langfuse when
Choose Langfuse when self-hosting matters, your stack is not LangChain-shaped, or you want the open-source core.
Choose LangSmith when
Choose LangSmith when your agents already run on LangChain or LangGraph and you want the evaluation workflow in the same house.
Sources and freshness
Compared from the verified profiles for Langfuse (checked 2026-10-07) and LangSmith (checked 2026-10-07). AgentsUse has not run a head-to-head benchmark, so this page claims no winner. It maps the differences your job has to price. Spotted an error? Send a correction.