Skip to main content

​
What We’ll Build

A research agent using LangGraph.js (TypeScript) that:
  1. Receives a topic as input
  2. Runs a researcher node to gather key facts
  3. Runs a reporter node to expand findings into a markdown report

​
Prerequisites

  • Node.js 20+
  • Crewship CLI installed and authenticated
Want to skip the setup? Clone the langgraph-js-quickstart repo and run crewship deploy to get a working agent deployed in minutes.

​
Step 1: Create the Project

mkdir research-agent-js && cd research-agent-js
mkdir src
Your project structure: 
research-agent-js/
├── src/
│   └── graph.ts
├── langgraph.json
├── package.json
├── tsconfig.json
└── crewship.toml

​
Step 2: Define Your Graph

Create src/graph.ts:
graph.ts
import { ChatOpenAI } from '@langchain/openai'
import { StateGraph, Annotation } from '@langchain/langgraph'

const State = Annotation.Root({
  topic: Annotation<string>(),
  research: Annotation<string>(),
  report: Annotation<string>(),
})

const llm = new ChatOpenAI({ model: 'gpt-4o-mini', temperature: 0.7 })

async function researcher(state: typeof State.State) {
  const response = await llm.invoke([
    {
      role: 'system',
      content:
        'You are a senior researcher. Given a topic, produce 10 concise bullet ' +
        'points covering the most important facts, recent developments, and key ' +
        'insights. Return only the bullet list.',
    },
    { role: 'user', content: `Topic: ${state.topic}` },
  ])

  return { research: typeof response.content === 'string' ? response.content : String(response.content) }
}

async function reporter(state: typeof State.State) {
  const response = await llm.invoke([
    {
      role: 'system',
      content:
        'You are a senior reporting analyst. Given research bullet points, ' +
        'expand them into a well-structured markdown report with an introduction, ' +
        'detailed sections, and a conclusion.',
    },
    { role: 'user', content: `Research notes:\n${state.research}` },
  ])

  return { report: typeof response.content === 'string' ? response.content : String(response.content) }
}

const builder = new StateGraph(State)
  .addNode('researcher', researcher)
  .addNode('reporter', reporter)
  .addEdge('__start__', 'researcher')
  .addEdge('researcher', 'reporter')
  .addEdge('reporter', '__end__')

export const graph = builder.compile()
The exported graph is what Crewship invokes. Your input (e.g., {"topic": "quantum computing"}) becomes the initial state.

​
Step 3: Add langgraph.json

Create langgraph.json in the project root — this lets crewship init auto-detect the framework:
langgraph.json
{
  "node_version": "20",
  "graphs": {
    "agent": "./src/graph.ts:graph"
  }
}

​
Step 4: Add Dependencies

Create package.json:
package.json
{
  "name": "research-agent-js",
  "version": "1.0.0",
  "type": "module",
  "dependencies": {
    "@langchain/langgraph": "^0.2.0",
    "@langchain/openai": "^0.3.0"
  }
}
Create tsconfig.json:
tsconfig.json
{
  "compilerOptions": {
    "target": "ES2022",
    "module": "ESNext",
    "moduleResolution": "bundler",
    "strict": true,
    "esModuleInterop": true
  }
}

​
Step 5: Add Crewship Configuration

Run crewship init to auto-generate the config — or create it manually:
crewship.toml
[deployment]
framework = "langgraph-js"
entrypoint = "./src/graph.ts:graph"
profile = "slim"

[build]
exclude = ["tests"]

[runtime]
timeout = 300
memory = 512
LangGraph.js uses a file path entrypoint (with ./), unlike Python frameworks which use a module path. The python field is not applicable.

​
Step 6: Set Environment Variables

crewship env set OPENAI_API_KEY=sk-proj-...

​
Step 7: Deploy

crewship deploy

📦 Packaging agent...
☁️  Uploading build context...
🔨 Building image...
✅ Deployed successfully!

Deployment: dep_abc123xyz
Project:    research-agent-js

​
Step 8: Run Your Agent

crewship invoke --input '{"topic": "quantum computing"}' --stream
Watch the execution: 
▶ Run started: run_xyz789
├─ [10:30:01] Starting graph execution
├─ [10:30:02] Node: researcher starting...
├─ [10:30:12] Node: researcher completed
├─ [10:30:13] Node: reporter starting...
├─ [10:30:45] Node: reporter completed
✅ Run completed in 44.8s

​
Step 9: Access the Output

The run result is the final graph state — an object with all state keys: 
curl https://api.crewship.dev/v1/runs/run_xyz789 \
  -H "Authorization: Bearer YOUR_API_KEY"

{
  "status": "succeeded",
  "result": {
    "topic": "quantum computing",
    "research": "• Quantum computers use qubits...",
    "report": "# Quantum Computing\n\n## Introduction\n..."
  }
}
To save the report as an artifact, write it to the artifacts/ directory inside your node:
import { writeFileSync, mkdirSync } from 'fs'

async function reporter(state: typeof State.State) {
  const response = await llm.invoke([...])
  const content = String(response.content)
  mkdirSync('artifacts', { recursive: true })
  writeFileSync('artifacts/report.md', content)
  return { report: content }
}

​
Next Steps

Streaming

Real-time event streaming from runs

Threads

Stateful conversations with LangGraph

Your First LangGraph Agent

Build the same agent in Python

Configuration Reference

crewship.toml options