LangfuseFreemium
An open-source (MIT core) observability and evaluation platform for LLM applications — tracing, prompt management, evals and cost monitoring. Acquired by ClickHouse in January 2026.
Analytics comparison
Pricing, pros, cons, and ideal use cases — side by side.
An open-source (MIT core) observability and evaluation platform for LLM applications — tracing, prompt management, evals and cost monitoring. Acquired by ClickHouse in January 2026.
An observability and evaluation platform for LLM and agent applications — tracing, datasets, automated evals, and production monitoring in one place.
| Langfuse | LangSmith | |
|---|---|---|
| Pricing | FreemiumCore is MIT-licensed and free to self-host. Langfuse Cloud: Hobby free (50,000 units, 2 users); Core $29/month, Pro $199/month and Enterprise $2,499/month, each with 100,000 units, then $8 per 100,000 up to 1M and $7 per 100,000 up to 10M. A unit is any trace, observation or score — including scores from Langfuse's own LLM-as-a-judge. As read on the vendor's pricing pages on 10 November 2026. | FreemiumDeveloper $0 (1 seat, 5,000 base traces a month). Plus $39 per seat/month with 10,000 base traces, then 0.005 LSU per base trace (1 LSU = $1, so $5 per 1,000) and 0.0025 LSU per trace upgraded to extended retention. Evaluator runs count as traces, and online evaluators upgrade the traces they score to extended retention by default. Enterprise is custom. As read on the vendor's pricing page and billing docs on 10 November 2026. |
| Category | Analytics | Analytics |
| Ideal for | Teams running LLM apps and agents in productionEnterprises needing self-hosted observabilityEngineering orgs standardizing AI tracing and evals | Teams running agents in productionEnterprises needing eval and monitoringEngineering orgs standardizing AI observability |
Langfuse is built around teams running llm apps and agents in production; LangSmith leans more toward teams running agents in production. Shortlist the one whose strengths line up with your biggest constraint.
Get one AI workflow a week showing Analytics in a real stack — what they cost, and where each one breaks.