LangChain
Open source agent frameworks plus the LangSmith platform for tracing, evaluating and deploying AI agents.
LangChain is a company and open source project for building AI agents. Developers use the LangChain, LangGraph and Deep Agents frameworks to write agents with any model provider, then use the LangSmith platform to trace runs, evaluate behaviour, deploy to production and monitor quality. A no-code product, Fleet, lets non-developers create agents in natural language.
Best for Engineering teams that build LLM agents in code and need tooling to test, deploy and observe them in production.
- Platforms
- Web, API, MCP
- Pricing model
- Freemium
- Free plan
- Yes
- API
- Yes
- Made by
- LangChain
- Added
Last verified: . Information taken from langchain.com.
What is LangChain?
LangChain is an agent engineering platform made up of open source frameworks and a commercial product called LangSmith. The frameworks cover three levels: langchain for getting an agent running quickly with a chosen model provider, langgraph for agents that need low-level control and some determinism, and deepagents for autonomous, long-running work. LangSmith follows the rest of the lifecycle. Observability gives step-by-step traces of agent runs, and Evaluation turns those traces into datasets and scores each step with online and offline evals. Deployment hosts long-running, stateful agents behind more than 30 Agent Server endpoints, while Sandboxes execute agent-generated code in isolation. An LLM Gateway applies spend limits, fallbacks and redaction of sensitive data to model calls, and Engine watches production traces, groups failures into issues and proposes fixes. Fleet offers agent building without code. LangSmith can run as managed cloud, in a hybrid setup or fully self-hosted.
Key Features
Open Source Frameworks
Provides langchain, langgraph and deepagents libraries for writing agents at different levels of control.
Observability
Records every step of an agent run so developers can debug it and query large traces quickly.
Evaluation
Builds datasets from real traces and runs online and offline evals, with annotation queues for human feedback.
Deployment and Sandboxes
Hosts agents on infrastructure built for persistent state and runs generated code in isolated, short-lived environments.
LLM Gateway
Sits between agents and model providers to control cost, rate limits, fallbacks and PII or secrets redaction.
Engine and Fleet
Engine detects and diagnoses production failures automatically; Fleet lets teams build agents in everyday language.
Screenshots
Videos
Use Cases
- Customer Support Agents
- Build and monitor an assistant that resolves support cases.
- Debugging an Agent
- Trace a failing run step by step to see which tool call or prompt went wrong.
- Regression Testing
- Evaluate a new prompt or model against a dataset before releasing it.
- Production Monitoring
- Track cost, latency, errors and quality across live agent runs with dashboards and alerts.
- Company-Wide Agents
- Give non-technical teams templates for agents that handle daily tasks.
Things to know
- Pricing page covers LangSmith
- The frameworks are open source; the paid plans and seat prices on the site apply to the LangSmith platform.
- Usage-based billing
- Traces beyond the included monthly amount, deployments and sandbox time are billed pay-as-you-go.
- Seat limits
- The Developer plan is limited to one seat, while Plus allows unlimited seats.
- Startup programme
- Startups with venture funding can apply for reduced pricing and usage credits.
FAQs
What is LangChain?
LangChain is an agent engineering platform that combines open source frameworks for building AI agents with LangSmith, a hosted platform for testing, deploying and monitoring them.
What is LangChain used for?
LangChain is used to build LLM-powered agents and applications, and its LangSmith platform is used to trace, evaluate and run those agents in production.
Is LangChain free?
LangChain's frameworks are open source, and LangSmith has a Developer plan at no seat cost for one user with a monthly allowance of traces, followed by pay-as-you-go usage.
How much does LangSmith cost?
LangSmith has Developer, Plus and Enterprise plans, with usage beyond the included amounts metered in LangChain Standard Units. See the official pricing page for current rates.
What is the difference between LangChain, LangGraph and Deep Agents?
LangChain is for starting agents quickly from templates, LangGraph is for agents that need low-level control, and Deep Agents is for highly autonomous, long-running tasks.
Can LangSmith be self-hosted?
Yes. LangSmith is offered as managed cloud, and the Enterprise plan adds hybrid and fully self-hosted options.
Does LangChain work with any model provider?
Yes. LangChain says its frameworks let developers build agents with any model provider and keep control of model choice.
Who it's for
- AI engineers
- Software developers
- Machine learning teams
- Startups
- Enterprises
Comments
No comments yet. If you have used LangChain, tell other readers how it went.
Alternatives to LangChain
| Tool | Free plan | Starting price | API |
|---|---|---|---|
 VercelCloud platform for deploying web apps and AI agents, with an AI gateway, sandboxes, workflows and a global CDN. |
Yes | $20/mo | Yes |
 FirecrawlWeb data API that searches, scrapes, crawls and interacts with websites and returns clean output for AI applications. |
Yes | $16/mo | Yes |
 PostHogProduct data platform combining analytics, session replay, feature flags, error tracking and a warehouse, with AI on top. |
Yes | — | Yes |
 DailyDeveloper platform for real-time voice, video and AI agents, built on WebRTC infrastructure and the open-source Pipecat framework. |
Yes | — | Yes |
 ModalServerless cloud platform for running AI inference, training, sandboxes and batch jobs on GPUs. |
Yes | — | Yes |
Listed in the same categories as LangChain. A starting price appears only where we checked it on the maker’s pricing page; a dash means not checked or not published. All AI Agents & Automation tools