Langgraph

LangGraph Plugin

LLM-native graph orchestration for complex stateful workflows.

Features

• State Machine Workflows - Define complex flows as graphs
• Durable Execution - Persist state across runs
• Human-in-the-Loop - Interrupt for approval/input
• Conditional Branching - Dynamic flow control
• Checkpointing - Save and restore state
• Streaming - Real-time node outputs

Quick Start

# Build complete LangGraph workflow
/langgraph:build my-workflow

# Add specific features
/langgraph:add node           # Add a graph node
/langgraph:add conditional    # Add conditional edge
/langgraph:add checkpoint     # Add state persistence
/langgraph:add human-loop     # Add human interruption

When to Use LangGraph

Use LangGraph when you need:

• Multi-step reasoning with state
• Human approval in AI workflows
• Retry/branching logic
• Persistent conversation state
• Complex agent orchestration

Agents

• graph-architect - Workflow design and structure
• node-specialist - Node implementation
• state-specialist - State management and persistence

Documentation

• LangGraph Docs
• LangGraph JS