LangChain Labs

Four offline arcs — Runnable contracts, toy RAG with citations, a human-in-the-loop state machine, and a stitched capstone assistant — each runnable with uv, each test-backed so you cannot lie with confidence.

First Runnable

lab-01-first-runnable

# lab-01-first-runnable — foundations arc Before you touch LangChain proper, internalize what a **Runnable** is supposed to do: same method names, predictable shapes, batching that is not "a for-loop

Rag Retriever

lab-02-rag-retriever

# lab-02-rag-retriever — RAG arc Retrieval without a vector database is still retrieval. You wire a toy corpus, score it with deterministic token overlap, simulate a reranker that breaks ties like gr

Hitl Stategraph

lab-03-hitl-stategraph

# lab-03-hitl-stategraph — LangGraph / HITL arc LangGraph nodes are cool; **interrupt/resume** is where production either earns trust or earns an incident. Here you simulate a tiny graph with an expl

Production Assistant

lab-04-production-assistant

# lab-04-production-assistant — capstone arc Wire the boring adult stuff together: **routing**, **retrieval**, a **human approval** path for spicy intents, a **trace log** you can actually read, and