From $49 per month🇺🇸
Prompt registry from New York for editing and versioning prompts without a code release, with evaluations and monitoring.
LangSmith AI features, pricing and GDPR (October 2026)
Prompt and Context Hub with versions, playground and AI chat, plus tracing and evals for LLM agents, in the cloud (US, EU) or self-hosted.
Per userFree plan
from $39 per user/month
Pricing details
According to the pricing page, Developer costs $0 per seat per month (1 seat, up to 5k base traces per month) and Plus costs $39 per seat per month (up to 10k base traces), then pay as you go. Enterprise is custom. As of October 2026.
What is LangSmith?
LangSmith is the commercial platform from LangChain for building and improving LLM agents. Its Prompt & Context Hub stores and versions prompts, and tracing, monitoring and evaluation show how agents behave in production, for engineering teams and the domain experts who edit prompts.
The Prompt & Context Hub holds prompts and contexts, which are versioned bundles of instructions and tools. A Playground tests prompts against model providers. Around it sit tracing, dashboards, datasets, experiments and annotation queues, agent deployment, sandboxes, an LLM gateway and Fleet for no-code agents. SDKs cover Python, TypeScript, Go and Java, and managed cloud, bring-your-own-cloud and self-hosted setups exist. LangChain began as a project in late 2022, became a company in early 2023 and states that it works with 35% of the Fortune 500. A developer is needed for setup.
LangSmith AI features and use cases
- Versions prompts outside the codePrompts are saved as commits in the hub, tagged, and can trigger webhooks, for instance to sync with GitHub. Domain experts edit wording in the interface while engineers pull the approved version, so behavior changes without a full deploy.
- Drafts and edits prompts with AIA prompt canvas lets a language model rewrite a prompt, change its reading level or length, and show a diff between versions. Chat inside the Playground can also help generate tools and output schemas for a prompt.
- Scores outputs with LLM judgesOnline evaluators grade production traces with an LLM as judge or with code, and dashboards alert by webhook or PagerDuty when quality or cost drifts. Teams notice regressions in agent answers without reading traces by hand.
- Clusters traces to find failure patternsInsights analyzes traces and groups them by topic to show usage patterns, common agent behaviors and failure modes, and it offers templates for error analysis. Product and engineering teams see where agents fail most often.
What sets LangSmith apart
- Combines a prompt hub with a context hub, where contexts are versioned bundles of instructions and tools promoted through environments.
- Cloud instances run in US, EU and Asia-Pacific regions, with bring-your-own-cloud and self-hosted options for stricter data residency needs.
- Works with any framework through SDKs for Python, TypeScript, Go and Java and through OpenTelemetry, not only with LangChain code.
- Seat plus usage pricing: one free Developer seat, then $39 per seat for Plus, with traces and other services metered.
Where LangSmith is strongest
LangChain, the company behind LangSmith, states that it is headquartered in San Francisco, with offices in New York, Boston and Amsterdam. That documents the United States as its home market, and the cloud also runs in EU and Asia-Pacific regions.
Questions to ask before you buy LangSmith
Pricing mixes seats with metered usage in LangChain Standard Units, so estimate trace volume with the vendor's calculator. The vendor says it has no EU legal entity for customer contracts, and a DPA is requested through support. Data location depends on the instance: Google Cloud in the US, EU or Asia-Pacific, or AWS in the US. Traces can contain personal data, so apply US state privacy laws, GDPR and HIPAA as relevant.
Using LangSmith in the EU: GDPR and AI Act
- Data locationMultiple regions, EU region available (according to the vendor). Hosting regions: United States, EU.
- Data processing agreementOn request.
- EU AI ActCheck whether the AI prepares decisions about people. If not, the high-risk rules usually do not apply. The high-risk obligations apply from December 2, 2027 (Digital Omnibus).
General information with sources, not legal advice.
LangSmith reviews
No reviews yet. Be the first to share how your team uses it.
Every review shows its verification level. We read each one before it goes live. How we check reviews
Questions about LangSmith
Is there a free plan?
Yes. The Developer plan costs $0 per seat per month, includes 1 seat and up to 5k base traces per month, and then bills pay as you go. Plus costs $39 per seat per month with up to 10k base traces. Prices as of October 2026.
Where is the data stored?
Organizations on smith.langchain.com are in Google Cloud US, those on eu.smith.langchain.com in Google Cloud EU, those on apac.smith.langchain.com in Google Cloud Asia-Pacific, and aws.smith.langchain.com is AWS US. Regional instances are available on all plans, and pricing is the same across regions.
Can LangSmith be self-hosted?
Yes. The vendor offers managed cloud, bring-your-own-cloud and self-hosted options for teams with data residency requirements and asks customers to contact it about the right setup. The pricing page lists self-hosted and hybrid deployment under Enterprise.
LangSmith update historyno updates yet
No changes recorded for LangSmith yet.
Changes the vendor made and confirmed user reports, each checked against the vendor's own pages. We review every tool monthly.
Alternatives to LangSmith
Tools in the same category (Versioning) with the most overlapping markets.
From $50 per month · Free plan🇺🇸
Open-source platform from Arize AI for tracing and evaluating LLM applications based on OpenTelemetry, self-hosted.
From $19 per month · Free trial🇺🇸
Open-source LLM observability and evaluation platform from Comet with prompt library, versions, LLM-as-judge metrics and prompt optimizer.
From $9 per user/month · Free plan🇺🇸
Prompt platform with a community library, Git-style versioning, testing across several models and an AI prompt generator.