Home

Multi-Agent Workflows using LangGraph — Project Wiki

What this repository implements: 18 runnable LangGraph exercises: graph nodes, edges, routing, state schemas, reducers, messages and composable subgraphs. The HR/IT examples demonstrate modular workflow handoffs; this is not yet an autonomous multi-agent system. An autonomous agent will be implemented in the next iteration with its own GPT-style small LLM.

Start here: Installation and Quickstart · Graph Examples · System Architecture · Low-Level Design.

Architecture at a glance

flowchart TB
    U["Learner / Developer"] --> CLI["Python CLI: argparse"]
    U --> NB["Jupyter Notebook"]
    CLI --> SELECT["Demo selection: --topic / --self-test"]
    NB --> EX["Notebook exercise cells"]
    SELECT --> EXAMPLES["18 independent example functions"]
    EX --> EXAMPLES
    subgraph LG["LangGraph orchestration"]
      BUILDER["StateGraph builder"] --> ROUTE["Nodes, edges and conditional routers"]
      ROUTE --> RUNTIME["Compiled graph: invoke"]
      RUNTIME <--> STATE["Typed state / reducers / messages"]
      RUNTIME --> SUB["Optional nested HR / IT subgraphs"]
    end
    EXAMPLES --> BUILDER
    EXAMPLES --> MSG["add_messages helper (no graph)"]
    EXAMPLES -. "2 opt-in topics" .-> LLM["ChatOpenAI / hosted OpenAI API"]
    RUNTIME --> OUTPUT["Printed JSON / optional results.json"]
    BUILDER --> MMD["Optional per-graph Mermaid .mmd export"]

Loading

Key capabilities

Area	What is implemented
Graph orchestration	StateGraph, nodes, normal and conditional edges, compile(), invoke()
Typed state	TypedDict, Pydantic, input/output/private state schemas
State reducers	Replace, operator.add numeric/list, custom merge, add_messages
Conversational state	MessagesState; offline canned reply, optional ChatOpenAI calls
Composition	HR→IT wrapper with state mapping; compiled IT subgraph with shared state
Developer tooling	18 CLI topics, offline self-test, JSON and Mermaid export, notebook

Guided reading

1. Repository Layout and Installation
2. System Architecture and Runtime Execution Flow
3. Low-Level Design, State Schemas, Reducers
4. Conditional Routing, Wrapper Subgraph, Direct Subgraph
5. LLM Integration, Testing, Security

Code links

• Python CLI implementation
• Runnable Jupyter notebook
• Dependency manifest