Sold by: LangChain
Deployed on AWS
LangSmith provides tools for developing, debugging, and deploying LLM applications. It helps you trace requests, evaluate outputs, test prompts, and manage deployments in one place.
LangSmith is framework agnostic, so you can use it with or without LangChain open-source libraries langchain and langgraph.
Prototype locally, then move to production with integrated monitoring and evaluation to build more reliable AI systems.
LangSmith provides:
- Observability to see exactly how your agent thinks and acts with detailed tracing and aggregate trend metrics.
- Evaluation to test and score agent behavior on production data and offline datasets for continuous improvement.
- Deployment to ship your agent in one click, using scalable infrastructure built for long-running tasks.
4.5
Overview
LangSmith Observability and Evals is a unified observability & evals platform where teams can debug, test, and monitor AI app performance - whether building with LangChain or not.
Find failures fast with agent observability. Quickly debug and understand non-deterministic LLM app behavior with tracing. See what your agent is doing step by step, then fix issues to improve latency and response quality.
Evaluate your agent's performance. Evaluate your app by saving production traces to datasets, then score performance with LLM-as-Judge evaluators. Gather human feedback from subject-matter experts to assess response relevance, correctness, harmfulness, and other criteria.
Experiment with models and prompts in the Playground, and compare outputs across different prompt versions. Any teammate can use the Prompt Canvas UI to directly recommend and improve prompts.
Track business-critical metrics like costs, latency, and response quality with live dashboards, then get alerted when problems arise and drill into root cause.
LangSmith Deployments is a purpose-built infrastructure and management layer for deploying and scaling long-running, stateful agents -- offering:
- 1-click deployment to go live in minutes,
- 30 API endpoints for designing custom user experiences that fit any interaction pattern
- Horizontal scaling to handle bursty, long-running traffic
- A persistence layer to support memory, conversational history, and async collaboration with human-in-the-loop or multi-agent workflows
- Native LangSmith Studio, the agent IDE, for easy debugging, visibility, and iteration
LangSmith Agent Builder: Give every team the ability to build, use, and improve AI agents with the security your org requires.
Highlights
- LangSmith Observability and Evals is a unified observability & evals platform where teams can debug, test, and monitor AI app performance - whether building with LangChain or not. Quickly debug and understand non-deterministic LLM app behavior with tracing. See what your agent is doing step by step, then fix issues to improve latency and response quality.
- LangSmith Deployments is a purpose-built infrastructure and management layer for deploying and scaling long-running, stateful agents offering 1/1-click deployment to go live in minutes, 2/Horizontal scaling to handle bursty, long-running traffic 3/A persistence layer to support memory, conversational history, and async collaboration with human-in-the-loop or multi-agent workflows.
- Please note: there is a $150k annual Platform License plus a minimum $150k annual usage commitment to access this package. To discuss enterprise pricing or to activate your commitment and obtain your license key after signup, please contact us at https://www.langchain.com/contact-sales - alternatively, our self-serve cloud-based products are available at https://www.langchain.com
Details
Sold by
Categories
Delivery method
Supported services
Delivery option
LangSmith Helm Chart
Latest version
Operating system
Linux
Deployed on AWS
New
Introducing multi-product solutions
You can now purchase comprehensive solutions tailored to use cases and industries.
Features and programs
Trust Center
Access real-time vendor security and compliance information through their Trust Center powered by Drata or Vanta. Review certifications and security standards before purchase.
Financing for AWS Marketplace purchases
AWS Marketplace now accepts line of credit payments through the PNC Vendor Finance program. This program is available to select AWS customers in the US, excluding NV, NC, ND, TN, & VT.
Pricing
Pricing is based on actual usage, with charges varying according to how much you consume. Subscriptions have no end date and may be canceled any time.
Additional AWS infrastructure costs may apply. Use the AWS Pricing Calculator to estimate your infrastructure costs.
Dimension | Cost/unit |
|---|---|
Per Trace | $0.01 |
Per Agent Run | $0.01 |
Metered Usage Amount | $0.01 |
Minimum annual usage commitment, billed in advance | $150,000.00 |
Per Agent Builder Run | $0.10 |
Dimensions summary
You pay for what you use across five independent metering dimensions. Per Trace charges cover observability data captured when your agents run. Per Agent Run and Per Agent Builder Run charge for agent executions on the platform. Metered Usage Amount tracks normalized compute and storage consumption from deployments, engine analysis, and related services. The Minimum annual usage commitment is billed in advance, setting a baseline you draw down against as usage accrues. These dimensions add together based on actual consumption, so your total scales with trace volume, agent activity, and resources used.
Top-of-mind questions for buyers
A trace is a single execution of your application—an agent run, evaluator, or playground session. One trace can include many steps, such as model calls and other tracked events. All those steps roll up into the single trace you are billed for.
It tracks normalized units of work and storage across services. Compute-related work is measured in LangChain Compute Units, and data stored or managed is measured in LangChain Storage Units. Deployments, engine analysis, and sandboxes all consume these units at different rates based on the resources they use.
It depends on your workload. High trace volume from heavy observability pushes Per Trace charges up. Running agents in production drives Per Agent Run and Metered Usage Amount through deployment uptime and engine analysis. All dimensions bill independently and add together on one invoice based on actual consumption.
www.langchain.com
Helpful?
Vendor refund policy
https://www.langchain.com/terms-of-service#:~:text=Customer%20will%20pay%20LangChain%20all ,Fees%20paid%20are%20non%2Drefundable.
Custom pricing options
Request a private offer to receive a custom quote.
Legal
Vendor terms and conditions
Upon subscribing to this product, you must acknowledge and agree to the terms and conditions outlined in the vendor's End User License Agreement (EULA) .
Content disclaimer
Vendors are responsible for their product descriptions and other product content. AWS does not warrant that vendors' product descriptions or other product content are accurate, complete, reliable, current, or error-free.
Delivery details
LangSmith Helm Chart
Supported services: Learn more
- Amazon EKS
Helm chart
Helm charts are Kubernetes YAML manifests combined into a single package that can be installed on Kubernetes clusters. The containerized application is deployed on a cluster by running a single Helm install command to install the seller-provided Helm chart.
Version release notes
LangSmith 0.13.14 release
Additional details
Usage instructions
See https://docs.smith.langchain.com/self_hosting for full installation and configuration instructions.
Resources
Vendor resources
Support
Vendor support
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.
Similar products
LangSmith is a framework-agnostic platform for building production AI agents. Trace, evaluate, and deploy agents with integrated observability and one-click deployment.
LangSmith BYOC deploys a fully managed agent observability, evaluation, and deployment platform inside your own AWS account and VPC, keeping sensitive data in your cloud boundary.
Enable your software teams to orchestrate AI agents across the entire SDLC with GitLab Duo Agent Platform Self-Hosted for offline licensing customers.
A production-ready foundation for deploying and governing AI agents on AWS. We took what we learned from building agent platforms for our clients and made it easy to deploy.
SAMP 2.0 is built on top of Amazon Bedrock AgentCore for single-agent and multi-agent systems. It helps platform teams standardize agent onboarding, bring governance and security from day one, and accelerate time-to-value.
No licenses, no subscription fees. You own the solution.
Customer reviews
Jai Y.
Eases AI Pipeline Building, Needs Better Debugging
Reviewed on Sep 29, 2026
Review provided by G2
What do you like best about the product?
I use LangChain to create agent and workflow chains. I find its modular architecture and standard interfaces really useful as they make it incredibly easy to string together LLMs, vector data, and memory into scalable, production-ready AI pipelines. With its modular design, I can swap models or databases with a single line of code without having to rewrite my app. It's very efficient in providing me with reusable session handling, prompt, output, input tool calling, so I don't have to create these things manually when creating an agent. Also, the initial setup was fairly easy for someone with knowledge about agent frameworks.
What do you dislike about the product?
LangChain's complex abstractions hide the underlying logic, making it difficult to debug, customize, and maintain as your project grows. It could simplify its abstractions, improve transparency into execution flow, and provide better debugging tools so developers can easily customize and troubleshoot complex workflows.
What problems is the product solving and how is that benefiting you?
Langchain provides reusable session handling and standard interfaces, so I don't have to manually create them, making it easy to build scalable AI pipelines.
Anonymous
Efficient Data Management with Slight Command Complexity
Reviewed on Sep 29, 2026
Review provided by G2
What do you like best about the product?
Langchain is great because it saves all my conversations and data, including large documents. I often forget to save my documents, but with Langchain, I just need to give a command, and it retrieves the document for me. The initial setup was easy, which I appreciated.
What do you dislike about the product?
Sometimes it's complicated to give commands because it adds extra layers just for one command and response feature.
What problems is the product solving and how is that benefiting you?
Langchain creates an efficient chatbot for me, saving all conversations, data, and PDFs. I can quickly access documents with a command, which is really helpful when I forget to store them.
Pharmaceuticals
Saves Development Time with Flexible LLM Support and Fast RAG Pipelines
Reviewed on Sep 28, 2026
Review provided by G2
What do you like best about the product?
makes it easy to connect different components together. The prompt templates, memory management, retrievers, vector databases integrations, & agent framework save a lot of development time. I especially like the support for multiple LLM providers, which makes it easy to switch between OpenAI, Anthropic, Azure OpenAI, & others without having to rewrite large portions of code.
Another feature I found really useful is the document processing pipeline. Loading PDFs, chunking documents, generating embeddings, & building a RAG application can be done fairly quickly compared to building everything from scratch.
Another feature I found really useful is the document processing pipeline. Loading PDFs, chunking documents, generating embeddings, & building a RAG application can be done fairly quickly compared to building everything from scratch.
What do you dislike about the product?
Sometimes examples from a few months ago no longer work because APIs have changed or modules have been deprecated. Documentation has improved but there are still occasions where I needed to go through GitHub issues or community posts to figure things out. There is no natural or automated way to automatically update the API if the former has been updated.
What problems is the product solving and how is that benefiting you?
For me, the biggest benefit has been in building RAG applications & AI assistants. It makes it much easier to connect LLMs with company documents, vector databases, APIs, & any external tools. Instead of spending weeks building the infrastructure, I can focus more on the actual business problem I'm trying to solve.
Izzy H.
Flexible AI Experimentation, but a Steep Learning Curve and Tricky Troubleshooting
Reviewed on Sep 14, 2026
Review provided by G2
What do you like best about the product?
What I like best about LangChain is the flexibility it gives me to experiment with AI without feeling locked into one specific setup. I can connect different models, data sources, and tools and see how they work together for a particular use case.
What do you dislike about the product?
The biggest challenge for me with LangChain is the learning curve. There are a lot of concepts, components, and different ways to approach the same task, so it can feel overwhelming when you're still getting familiar with the platform.
I've also found that troubleshooting isn't always straightforward. Sometimes a workflow that seems fairly simple can require more configuration than expected, and when something doesn't work, it can take time to figure out exactly where the issue is coming from.
I've also found that troubleshooting isn't always straightforward. Sometimes a workflow that seems fairly simple can require more configuration than expected, and when something doesn't work, it can take time to figure out exactly where the issue is coming from.
What problems is the product solving and how is that benefiting you?
LangChain helps me bridge the gap between having an idea for an AI solution and actually testing how it could work in practice. Instead of working with an AI model in isolation, I can connect it to different tools, information sources, and processes to create something more useful.
Deepak A.
Easy Model Switching Across Providers
Reviewed on Sep 04, 2026
Review provided by G2
What do you like best about the product?
I can use it with any model provider. Switching between models is easy.
What do you dislike about the product?
I find it very unstable, with every new release my project breaks.
What problems is the product solving and how is that benefiting you?
I’m using it to build an autonomous API testing framework that can detect backend APIs and generate automation code within the automation framework.