LangGraph Samples

[mfateev]

Summary

Add LangGraph samples demonstrating the Temporal LangGraph integration for durable AI agent workflows.

Samples Included

Sample	Description
hello_world	Minimal starter example - basic plugin setup and graph registration
react_agent	ReAct agent with tool calling and multi-step reasoning
approval_workflow	Human-in-the-loop with interrupt/resume for approvals
supervisor	Multi-agent supervisor pattern coordinating specialized agents
agentic_rag	Retrieval-augmented generation with document grading
deep_research	Multi-step research with web search and iterative refinement
plan_and_execute	Plan-and-execute pattern with structured step execution
reflection	Self-reflection pattern for iterative improvement

Key Features Demonstrated

• Durable execution: Each graph node runs as a Temporal activity with automatic retries
• Human-in-the-loop: Interrupt/resume workflows for human approval
• Multi-agent coordination: Supervisor patterns for agent orchestration
• RAG patterns: Document retrieval, grading, and query rewriting
• Planning patterns: Structured plan-and-execute workflows

Prerequisites

• Temporal server running locally
• Python 3.9+
• OpenAI API key (for LLM-based samples)

Running

# Install dependencies
uv sync --group langgraph

# Install SDK from langgraph-plugin branch
uv pip install "temporalio @ git+https://github.com/mfateev/sdk-python.git@langgraph-plugin"

# Run a sample (e.g., hello_world)
uv run langgraph_samples/hello_world/run_worker.py  # Terminal 1
uv run langgraph_samples/hello_world/run_workflow.py  # Terminal 2

Related PR

• SDK: https://github.com/temporalio/sdk-python/pull/1263