LangChain Launches LangGraph Cloud — Managed Infrastructure for Stateful AI Agents
LangChain has launched LangGraph Cloud, a managed platform for deploying stateful multi-agent applications. The service handles agent persistence, fault tolerance, human-in-the-loop workflows, and horizontal scaling out of the box.
LangChain announced the general availability of LangGraph Cloud, a managed deployment platform specifically designed for stateful AI agent applications built with LangGraph.
While LangGraph (the open-source framework) handles the agent logic, LangGraph Cloud provides the production infrastructure: persistent state management, automatic checkpointing, fault recovery, horizontal scaling, and built-in human-in-the-loop capabilities.
Core Features:
- •Persistent Agent State: Agent conversations and workflow state survive server restarts and can be resumed at any point.
- •Time Travel Debugging: Rewind agent execution to any previous state, modify inputs, and replay — invaluable for debugging complex agent behaviors.
- •Cron Agents: Schedule agents to run periodically for monitoring, reporting, or maintenance tasks.
- •Streaming Support: Real-time streaming of agent thoughts, tool calls, and outputs to frontend clients.
- •Multi-Tenant: Built-in support for serving multiple users/organizations with isolated agent state.
Pricing starts at $0 for development, with production plans based on compute and storage usage.
"We've seen thousands of teams build incredible agent prototypes with LangGraph, but deploying them to production was still too hard," said Harrison Chase, CEO of LangChain. "LangGraph Cloud eliminates that gap."