LangSmith
Tracing, evaluation and prompt management for LLM applications, from the LangChain team.
why this verdict
Keep paying — OpenAI has not replaced it
OpenAI's version overlaps, but does not finish the dev tools job, so this one still earns its $39/mo.
- A named launch, not a vibe named, unlinked
- Model providers shipping free tracing and evals inside their own developer consoles. OpenAI, August 5, 2025 — no announcement link recorded yet.
- How much of the job it covers not the job editorial call
- Parts of it. The job still needs the tool to get finished.
- Is there a free way to do it? yes
- 2 of 3 listed replacements have a usable free tier: Langfuse and Helicone.
- What the call is worth $468/yr
- $39/mo at the entry paid tier — $468 a year per seat.
- Threatened by
- OpenAI
- Since
- August 5, 2025
- List price
- $39/mo
- Per year
- $468
The backstory
LangSmith is the observability layer for LLM apps: every chain, tool call and token accounted for, with datasets and evaluators to catch regressions before a prompt change ships. Model vendors now include basic tracing free, which caps the entry-level. LangSmith stays relevant because serious teams run multiple providers and need one place to compare them, and because evaluation, not logging, is the hard part it invested in most heavily.
Escape hatches
Open-source and self-hostable with a similar feature set
langfuse.com open_in_newEvaluation-first workflow with strong dataset tooling
braintrust.dev open_in_newLightweight proxy-based logging and cost tracking
helicone.ai open_in_new2 of 3 replacements have a usable free tier.
