Overview
LangSmith Observability
LangSmith Observability gives you complete visibility into agent behavior.
LangSmith Observability
LangSmith Deployments
LangSmith Evaluations
LangSmith Fleet
Agent Development Lifecycle

Product video
LangSmith Agent Engineering Platform - Debug, Evaluate, and Deploy AI Agents with Confidence
LangSmith is a unified observability, evaluation, and deployment platform built for teams shipping AI agents to production - whether you build with LangChain, LangGraph, or any other framework. Backed by Sequoia Capital and recognized on the Forbes AI 50 list in April 2025, LangChain provides the infrastructure that engineering teams need to move from prototype to production-grade AI.
Find and Fix Agent Failures Fast
Non-deterministic LLM behavior makes debugging difficult. LangSmith tracing gives you step-by-step visibility into every action your agent takes, so you can pinpoint failures, reduce latency, and improve response quality. Stop guessing why your agent misbehaved - see exactly what happened and fix it.
Evaluate Agent Performance Systematically
- Automated evaluation: Save production traces to datasets and score performance with LLM-as-Judge evaluators to catch regressions before your users do
- Human feedback: Gather assessments from subject-matter experts on response relevance, correctness, harmfulness, and other domain-specific criteria
- Prompt experimentation: Test models and prompts side by side in the Playground, compare outputs across prompt versions, and let any teammate use the Prompt Canvas UI to recommend improvements
Monitor Production Health in Real Time
Track business-critical metrics including costs, latency, and response quality through live dashboards. Set alerts for anomalies and drill into root cause when problems arise - giving your team the operational confidence to run AI agents at scale.
Deploy and Scale Stateful Agents with LangSmith Deployments
LangSmith Deployments provides purpose-built infrastructure for running long-lived, stateful agents in production:
- 1-click deployment to go live in minutes without managing infrastructure
- 30 API endpoints for designing custom user experiences that fit any interaction pattern
- Horizontal scaling to handle bursty, long-running agent traffic reliably
- Built-in persistence layer supporting memory, conversational history, and async collaboration for human-in-the-loop and multi-agent workflows
- LangSmith Studio - an agent IDE for debugging, visibility, and rapid iteration on deployed agents
Example Workflow: Improving a Customer-Facing AI Agent
A team deploys a customer support agent that answers questions using internal knowledge bases. Using LangSmith tracing, they identify that the agent occasionally retrieves irrelevant documents, leading to incorrect answers. They save these failure traces to a dataset, run LLM-as-Judge evaluators to score correctness at scale, and gather human feedback from domain experts. After iterating on prompts in the Playground, they redeploy with 1-click and monitor the improved agent's performance on live dashboards - closing the loop from detection to resolution.
Built on a Proven Open-Source Foundation
LangChain launched as an open-source project in October 2022 and has grown into one of the most widely adopted frameworks for building LLM-powered applications. LangSmith extends this foundation with production-grade tooling for observability, evaluation, and deployment. The platform is available as a fully managed SaaS through AWS Marketplace with metered billing. Visit the Pricing tab for details on billing dimensions and plan options.
To get started, subscribe through AWS Marketplace and visit our support portal for onboarding guidance.
Highlights
- End-to-end observability for AI agents: LangSmith provides detailed tracing that shows exactly how your agent thinks and acts at every step. Track costs, latency, and response quality with aggregate trend metrics and dashboards. Debug failures fast by inspecting full request traces across your entire agent workflow - whether you build with LangChain, LangGraph, or any other framework. LangSmith is fully framework agnostic, so it integrates into your existing stack without requiring migration.
- Built-in evaluation and testing for continuous improvement: Score and test agent behavior using LLM-as-Judge evaluators, custom scoring functions, or human review on both production data and offline datasets. Use the Prompt Canvas UI to experiment with prompt variations and compare outputs side by side. Run evaluations continuously to catch regressions before they reach users, helping teams build more reliable AI systems through data-driven iteration rather than guesswork.
- One-click agent deployment on scalable infrastructure: Ship agents to production in a single click using infrastructure purpose-built for long-running, stateful AI tasks. LangSmith Deployments provides horizontal scaling to handle bursty traffic, a persistence layer for human-in-the-loop workflows, and over 30 API endpoints for integration. Built on the LangGraph Platform, it supports cron jobs, streaming, and double-texting out of the box.
Details
Features and programs
Financing for AWS Marketplace purchases
Pricing
Dimension | Description | Cost/unit |
|---|---|---|
Usage Fee | Units of metered usage as calculated based on platform activity | $0.01 |
Dimensions summary
Top-of-mind questions for buyers
Vendor refund policy
Custom pricing options
How can we make this page better?
Compliance
Compliance profile
- Source
- Vanta Trust Center





Legal
Vendor terms and conditions
Content disclaimer
Delivery details
Software as a Service (SaaS)
SaaS delivers cloud-based software applications directly to customers over the internet. You can access these applications through a subscription model. You will pay recurring monthly usage fees through your AWS bill, while AWS handles deployment and infrastructure management, ensuring scalability, reliability, and seamless integration with other AWS services.
Resources
Vendor resources
Support
Vendor support
LangChain provides support for LangSmith customers through its support portal at https://support.langchain.com/ . Visit the portal to submit requests, report issues, or get help with platform functionality including tracing, evaluation, and deployment.
For additional information or to reach the LangChain team, please visit the support portal linked above.
AWS infrastructure support
AWS Support is a one-on-one, fast-response support channel that is staffed 24x7x365 with experienced and technical support engineers. The service helps customers of all sizes and technical abilities to successfully utilize the products and features provided by Amazon Web Services.