LangSmith
What is LangSmith?
LangSmith is an AI observability and evaluation platform for teams that need to trace, debug, score, and deploy agents. It combines Observability, Evaluation, Deployment, Prompt Hub, Playground, and Canvas, plus annotation queues for human feedback. LangSmith works with Python, Typescript, Go, or Java SDKs and is used by Klarna, Vanta, monday.com, and Cloudflare. Plans run Developer $0 / seat per month, Plus $39 / seat per month, and Enterprise custom pricing.
Last verifiedHow we evaluate
At a glance
- LangSmith is best for AI teams who need to trace, evaluate, and ship agents with production visibility.
- Developer $0 / seat /mo; Plus $39 / seat /mo; Enterprise Custom
What does LangSmith do?
LangSmith turns agent traces into a workflow for debugging, scoring, and shipping. It captures full conversations and agent runs, then layers Observability, Evaluation, and Deployment so teams can inspect every step, run LLM-as-judge or code-based checks, and push agents into production with a standardized path. The platform also includes Prompt Hub, Playground, and Canvas for prompt iteration, plus annotation queues for human feedback. At scale, LangSmith says it serves 35% of the Fortune 500, has crossed 1 billion open source downloads, and ingests over 1 billion events per day. It is framework agnostic and works with Python, Typescript, Go, or Java SDKs, so teams can trace their preferred stack without changing architecture. Customers including Klarna, Vanta, monday.com, and Cloudflare use it to understand and improve agents across production workloads.
Why use LangSmith?
- Framework-agnostic SDKs let teams instrument agents without rewriting their stack.
- Observability, evaluation, and deployment live together, so debugging and release workflows stay connected.
- Human feedback queues and prompt tools shorten the loop between trace review and improvement.
- Enterprise hosting options keep data inside your environment when compliance requires it.
- Fleet and sandboxes extend the platform beyond tracing into controlled agent execution and safe code runs.
Who is LangSmith for?
- AI platform teams who need a shared system for tracing and monitoring agent behavior.
- ML engineers who want production trace data to guide evals and prompt changes.
- Product teams shipping agents who need a standardized deployment path and feedback loops.
- Engineering leaders who need visibility into cost, latency, errors, and failure modes.
- Teams with strict data controls who need managed cloud, BYOC, or self-hosted options.
What are LangSmith's key features?
Observability
Track agent runs, traces, and costs across millions of traces with sub-second performance, so teams can spot failures and control spend.
Evaluation
Run offline and online evals with Prompt Hub, Playground, and Canvas to compare changes and improve prompts before shipping.
Deployment
Deploy agents with one click using 30+ API endpoints and one standard deployment pattern, reducing setup work for production releases.
Fleet
Manage Fleet agents and runs at scale, with up to 50 Fleet runs per month in the base plan and higher limits available.
Human Feedback
Collect annotation queue feedback from reviewers to label outputs and guide model improvements using real human judgments.
Monitoring
Set monitoring and alerting with Webhook and Pagerduty alerts, helping teams catch regressions and respond faster in production.
Agent-native observability & evals
Connect OpenTelemetry, MCP, A2A, and Agent Protocol to trace agent decisions and evaluate workflows built on open source frameworks.
What does LangSmith integrate with?
- OpenTelemetry
- A2A
- MCP
- Agent Protocol
- Pagerduty
- OpenAI
- Anthropic
- Gemini
- Salesforce
- Gmail
- Slack
What are LangSmith's use cases?
Platform tracing for AI teams
AI platform teams use LangSmith to trace agent behavior across production workflows, using Observability and Tracing to see where tool calls, prompts, and model outputs break down. They pair that with Monitoring to catch regressions early and keep agent systems reliable at scale.
Eval-driven prompt iteration
ML engineers use LangSmith to turn production trace data into better prompts and tests, using Evaluation and Offline Evals to compare changes before rollout. With Online Evals and Insights, they can validate whether a prompt update actually improves real user outcomes.
Standardized agent deployment
Product teams shipping agents use LangSmith to move from prototype to production with a repeatable path, using Deployment and Fleet to launch and manage agent workloads. They use Human Feedback to close the loop on user corrections and improve the next release.
Operational visibility for leaders
Engineering leaders use LangSmith to keep an eye on cost, latency, errors, and failure modes, using Monitoring and Agent-native observability & evals to understand what is happening in production. That visibility helps them prioritize fixes before agent issues affect customers.
How does LangSmith work?
- Connect your first model or agent workflow with Tracing so LangSmith can capture each step, prompt, tool call, and output in one timeline.
- Review traces in Observability and Insights to spot latency spikes, cost outliers, and failure patterns across real production runs.
- Create test cases in Evaluation, then run Offline Evals and Online Evals to compare prompt or model changes before and after release.
- Collect reviewer input with Human Feedback and annotation queues, then use those signals to refine prompts and reduce recurring errors.
- Deploy with Deployment and Fleet, then keep Monitoring and alerting on so you can catch regressions and roll out improvements safely.
How much does LangSmith cost?
Developer
$0 / seat per month- Up to 5k base traces / mo, then pay-as-you-go
- Tracing to debug agent execution
- Online and offline evals
- Prompt Hub, Playground and Canvas for auto improving prompts
- Annotation queues for human feedback
- Monitoring and alerting
- 1 Fleet agent
- Community support
- 1 seat
Plus
$39 / seat per month- Everything in the Developer plan, and:
- Up to 10k base traces / mo, then pay-as-you-go
- 1 dev-sized agent deployment included
- Email support
- Unlimited Fleet agents
- Add unlimited seats
- Up to 3 workspaces
Enterprise
Custom pricing- Everything in the Plus plan, and:
- Alternative hosting options, including hybrid and self-hosted so data doesn't leave your VPC
- Custom SSO and RBAC
- Acccess to deployed engineering team
- Support SLA
- Team trainings & architectural guidance
- Custom Fleet packages
Frequently asked questions
What is LangSmith?
LangSmith is an AI observability and evaluation platform for teams that need to trace, debug, score, and deploy agents. It combines Observability, Evaluation, Deployment, Prompt Hub, Playground, and Canvas, plus annotation queues for human feedback. LangSmith works with Python, Typescript, Go, or Java SDKs and is used by Klarna, Vanta, monday.com, and Cloudflare. Plans run Developer $0 / seat per month, Plus $39 / seat per month, and Enterprise custom pricing.
How much does LangSmith cost? Is it free?
LangSmith has a free plan, with paid tiers including Plus at $39 / seat per month, Enterprise at Custom pricing.
What is LangSmith used for? Who is it for?
LangSmith is used for Observability, Evaluation, and Deployment. It's built for AI platform teams, ML engineers, and Product teams shipping agents.
Does LangSmith have an API and what does it integrate with?
LangSmith doesn't publish a public API. It integrates with OpenTelemetry, A2A, MCP, Agent Protocol, Pagerduty, and 6 more.
Editor's read
Developer includes up to 5k base traces per month, while Plus raises that to 10k and adds unlimited seats. If your trace volume is near either ceiling, verify how quickly pay-as-you-go charges would begin.
Explore other Agent Tools & Integrations
