Langfuse vs LangSmith
Prices and features checked 2026-10-17 on both vendors' own pages. “not verified” means we could not confirm it.
Langfuse or LangSmith? Langfuse is open source (MIT core), free to self-host, and its cloud charges a flat fee plus units: Core is $29/month with unlimited users and 100k units, then $8 per 100k. LangSmith is LangChain's hosted platform, charged per seat: Plus is $39/seat/month with 10k base traces, and self-hosting is on Enterprise only. Both support OpenTelemetry, CI test runs, LLM-as-judge, annotation queues, prompt versioning and versioned datasets.
| Langfuse | LangSmith | |
|---|---|---|
| Made by | Langfuse (ClickHouse, Inc.) | LangChain |
| Licence | MIT for code outside the ee/ directories; ee/ code under a separate licence [source] | Hosted service; self-hosted and hybrid options on Enterprise [source] |
| Self-hosting | Free to self-host (Docker Compose or Kubernetes) [source] | Self-hosted or hybrid on Enterprise only [source] |
| Meter | Billable units: traces, observations (spans, events, generations) and scores [source] | Seats plus base traces (14-day retention) and extended traces (180-day) [source] |
| Free tier | Hobby: 50k units/month, 30 days data access, 2 users [source, checked 2026-10-17] | Developer $0: 1 seat, up to 5k base traces/month, then pay as you go [source, checked 2026-10-17] |
| Entry paid price | Core $29/month: 100k units included, then $8 per 100k units (graduated down to $6 per 100k above 50M) [source, checked 2026-10-17] | Plus $39/seat/month: up to 10k base traces/month, then pay as you go (per-trace rate not stated on the pricing page) [source, checked 2026-10-17] |
| Retention | Data access: 30 days (Hobby), 90 days (Core), 3 years (Pro and Enterprise) [source] | Base traces 14 days; extended traces 180 days (extra fee) [source] |
| OpenTelemetry | OpenTelemetry ingestion on all plans [source] | OpenTelemetry tracing via OTLP endpoint [source] |
| CI gating | Run experiments in CI and gate changes [source] | pytest and Vitest/Jest integrations that raise assertion errors in CI [source] |
| LLM-as-judge | LLM-as-judge evaluators on all plans [source] | LLM-as-a-judge evaluators [source] |
| Human annotation | Annotation queues: 1 on Hobby, 3 on Core, unlimited on Pro [source] | Single-run and pairwise annotation queues [source] |
| Prompt management | Prompt versioning and release labels [source] | Prompt versions, environments and access controls [source] |
| Dataset versioning | Dataset versions are accessible via the SDK/API [source] | Every change creates a dataset version; versions can be tagged [source] |
| SOC 2 / HIPAA | SOC 2 Type II and ISO 27001 reports and HIPAA (BAA) on Pro ($199/month) and Enterprise [source] | LangChain states it maintains SOC 2 Type II, GDPR and HIPAA compliance [source] |
Verdict: when to use which
Use Langfuse if you want to self-host for free, or your team is large enough that per-seat pricing adds up (Langfuse Core has unlimited users).
Use LangSmith if you build on LangChain or LangGraph and want tracing, evals and agent deployment from the same vendor, or you are a solo developer within the free Developer seat (5k base traces).
The meters differ, so price both on a week of your own traffic in the pricing calculator.
FAQ
Does Langfuse work with LangChain?
Yes: Langfuse lists native framework integrations, and both tools accept OpenTelemetry traces, so you are not tied to LangChain either way.
Which is cheaper for a 10-person team?
On list prices checked 17 Oct 2026, LangSmith Plus is $390/month in seats before any trace overage; Langfuse Core is $29/month plus usage above 100k units. Your trace volume decides the rest; use the pricing calculator.
Do both publish SOC 2 and HIPAA?
Langfuse lists SOC 2 Type II and ISO 27001 reports and HIPAA (BAA) on Pro and Enterprise. LangChain states it maintains SOC 2 Type II, GDPR and HIPAA compliance.
More
Other tool comparisons: Braintrust vs Langfuse · Braintrust vs LangSmith · Arize Phoenix vs Langfuse · Comet Opik vs Langfuse · Promptfoo vs DeepEval / Confident AI