







Three ways to build,
from full harness to full control
Each package solves a different problem. Choose whereyou want to start, then compose the rest ofthestack around it.
APIs to handle real-world agent interactions
Run human-in-the-loop approvals, background tasks, and multi-agent coordination on a durable runtime with built-in fault tolerance.
- Deploy with one click
- Production-ready APIs for HIL, memory, and conversation threading
- Stream type-safe messages, UI components, and custom events

Align agent operations, org wide
Manage agents through a centralized registry with versioning, instant rollbacks, and native A2A, MCP, and Agent Protocol support.

Scale without rebuilding infrastructure
Running agents gets harder at scale. Handle bursty, long-running, and complex agent workloads with horizontal scaling.




Infrastructure for production agents
Every agent in production faces the same needs: durable execution, streaming, state and task management. LangSmith Agent Server ships with all of it, so you spend your time on your agent logic instead of rebuilding the wheel.
Protocol endpoints
Centrally manage all agents with versioning
Share, modify, and reuse agents
Add custom auth and access control
Run agents on a schedule
Handle new input while an agent is running
Maintain context across restarts and cold-starts
Review, edit, and approve actions
Persistent checkpoints with up to 25 MB payloads
Semantic search for long-term memories
State isolation for subagents
Stream outputs in real time
Debug visually with our IDE
Manage rapid-fire requests
Trigger external systems
One-click deploy
Deploy your agent in one click and get durable execution, state, and streaming out of the box.
Resources for LangSmith Agent Server
FAQs for LangSmith Agent Server
Yes, LangSmith Deployment has been renamed to LangSmith Agent Server as of November 2026.
LangGraph is the open-source framework for building stateful, orchestrated agent workflows. LangSmith Agent Server is the managed service for running those agents at scale in production.
No — you can deploy agents built with any framework. Check out the docs to get started.
You can deploy your agent using LangSmith Agent Server, available in the LangSmith Plus and Enterprise plans. which has the following agent deployment options:
Cloud: Fully managed and hosted as part of LangSmith (our unified observability and evals platform). Deploy quickly, with automatic updates and zero maintenance.
Hybrid (SaaS control plane, self-hosted data plane). No data leaves your VPC. Provisioning and scaling is managed as a service.
Fully Self-Hosted: Deploy your agent entirely on your own infrastructure.
Ready to deploy your agent to production?
Run your agent with production-ready infrastructure. Get started in minutes with 1-click deployments, built-in APIs, and autoscaling to handle enterprise-scale traffic.



