How to Deploy LangGraph to Production

You built a LangGraph agent. It runs locally with nodes, edges, conditional routing, and stateful execution. Now you want to deploy it for others to use.

LangGraph adds complexity that plain LangChain chains don't have: state management, cycles, conditional edges, and long-running executions. That complexity matters in production.

The deployment challenges

State management — LangGraph graphs are stateful by design. Your StateGraph defines typed state that flows through nodes. Locally this lives in memory, but in production you need state that survives restarts, stays isolated between concurrent runs, and works across multiple machines.

Long-running executions — Graphs with conditional loops and multi-step reasoning can run for minutes. You need async execution, background processing, and streaming support to preserve LangGraph's event streaming capabilities.

Complex tracing — When a graph fails, debugging requires understanding which node threw the error, what the state was, whether conditional edges routed correctly, and if cycles ran as expected. Without proper observability, this becomes painful fast.

Add containerization, dependency management, scaling, versioning, auth, and rate limiting — you're building infrastructure instead of your product.

Deploy with Crewship

Crewship now supports LangGraph natively. If your project has a langgraph.json file, Crewship auto-detects it and handles everything.

Quick start

Install the CLI:

curl -fsSL https://www.crewship.dev/install.sh | bash

Log in and initialize:

crewship login
crewship init

Crewship detects LangGraph and generates a crewship.toml:

[deployment]
framework = "langgraph"
entrypoint = "src.my_graph.graph:graph"
python = "3.11"
profile = "slim"

The entrypoint points to your compiled StateGraph object. No API code needed.

Deploy:

crewship deploy

Your code gets packaged, containerized, and deployed. You get a deployment URL and console access.

Add secrets:

crewship env set OPENAI_API_KEY=sk-... TAVILY_API_KEY=...
# Or import from .env
crewship env import -f .env

Secrets are encrypted and injected at runtime.

Run your graph

From the CLI:

crewship invoke --input '{"messages": [{"role": "user", "content": "Research the latest AI papers"}]}'

The CLI streams events as nodes execute in real-time.

From the REST API:

curl -X POST https://api.crewship.dev/v1/runs \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"deployment": "my-graph", "input": {"messages": [{"role": "user", "content": "Research the latest AI papers"}]}}'

Execution traces map to your graph nodes — see which node ran, its state, output, and timing. When something fails, you'll see exactly where and why.

LangGraph.js support

Building with TypeScript/JavaScript? Crewship supports LangGraph.js with the same CLI workflow, API, and execution traces. Check out the LangGraph.js page for details.

What you get

All the infrastructure you'd otherwise build yourself:

• Isolated execution — every run gets its own environment

• Auto-scaling — scales up with demand, down to zero when idle

• Deployment versioning — each deploy creates a new version, roll back anytime

• Graph-aware traces — see nodes, state, timing, and token usage at each step

• Webhooks — trigger runs from CI/CD, cron, or Zapier; get completion notifications

• Token auth — API key authentication, manage keys from the console

• Real-time SSE streaming — watch execution live or poll for results

Get started

Deploy your first LangGraph graph in minutes. No credit card required for the free tier.

• Get started

• Read the docs

• LangGraph framework page

If you're already using Crewship for CrewAI, the same account and CLI work for LangGraph. Just crewship init in your LangGraph project and deploy.