langgraph-python

LangGraph Python + AMC

A LangGraph stateful agent with a planner→executor graph, routing all LLM calls through the AMC Gateway.

What It Does

• Builds a two-node graph: planner (breaks question into steps) → executor (provides final answer)
• Uses LangGraph's StateGraph with typed state and conditional edges
• AMC captures all LLM interactions transparently via env var proxy

Prerequisites

• Python ≥ 3.10
• AMC CLI installed with the checksum-verified GitHub Release installer (npm is not public yet; see the root README)
• An OpenAI API key

Install

cd examples/langgraph-python
pip install -r requirements.txt

Run with AMC

amc up
amc wrap langgraph-python -- python main.py

Expected Output

[AMC] Routing LLM calls through gateway: http://localhost:3700/v1
=== LangGraph Stateful Agent ===
Steps taken: 2
Final answer: The sky appears blue because...
[AMC] All LLM calls captured as evidence via gateway proxy.

How AMC Integrates

AMC sets OPENAI_BASE_URL to the gateway. LangGraph uses LangChain's ChatOpenAI under the hood, which reads env vars automatically. The graph structure, state transitions, and multi-step reasoning are all captured as evidence.

The adapter definition is in src/adapters/builtins/langgraphPython.ts.