AI Agent Quickstart
Build and run your first durable AI agent in minutes.
Why Restate?
Restate is the foundation for AI workflows, agents, and agentic platforms, written as plain code in your language of choice:- ✅ Durable Execution: Retry and recover from failures without repeating completed steps.
- ✅ State & session coordination: Built-in multi-turn, persistent sessions. Scale out to thousands of concurrent sessions with built-in concurrency control
- ✅ Communication & flow control: Route tasks between distributed agents, humans, and MCP servers, and shape traffic with concurrency control
- ✅ Observability & control: Inspect detailed execution journals, and pause, cancel, kill, or roll back stuck agents
- ✅ Production safety: Approvals, timeouts, rollbacks, and escalation without callback plumbing
SDK integrations
Any LLM SDK — write the agent loop yourself with OpenAI, Anthropic, Vercel AI SDK, Google Gen AI, LiteLLM, or any other SDK. Wrap LLM calls inctx.run() for durability.
Vercel AI SDK
LiteLLM
Any other SDK
Vercel AI SDK
OpenAI Agents SDK
Google ADK
Pydantic AI
LangChain
Integrating with other SDKs
Getting Started
The following pages give you a tour of how to build durable AI agents and workflows with Restate:1. Durable Agents
Implement agents that survive crashes and recover automatically. Every LLM call and tool execution is durably persisted.
2. Durable Sessions
Add persistent sessions keyed by ID with built-in concurrency control. Conversation state survives crashes and restarts.
3. Approvals with Pause & Resume
Add resilient human approvals that pause the agent and resume when the response arrives, even across restarts.
4. Multi-Agent Orchestration
Route tasks between specialized agents with durable routing decisions. Coordinate via handoffs, tools, or remote calls.
5. Observability & Control
Inspect agent execution step by step, export traces, and cancel or kill stuck agents.