martinpercu/langchain-langgraph_agents-structure

How martinpercu/langchain-langgraph_agents-structure works

This repository is a LangGraph/LangChain agent template. Its main user-to-result path is a FastAPI chat request invoking the support-agent graph, which extracts customer details, routes to conversation or booking, and returns a response; an alternate API setup adds PostgreSQL checkpointing. Separate agent modules demonstrate RAG, ReAct, and other workflow patterns. The support-agent graph and API wiring are shown in source excerpts; broader example capabilities are documented in the README, so their internal flows are not inferred.

Main entry points: support_agent.py Support Graph · extractor/node.py Customer Extractor · route.py Intent Router · conversation/node.py Conversation Node

Components 27

API and Persistence

Chat API src/api/main.py
Persistent Chat API src/api/main-2.py
Checkpoint Lifecycle src/api/db.py

Support Agent

Support Graph src/agents/support_agent/support_agent.py
Conversation State src/agents/support_agent/state.py
Customer Extractor src/agents/support_agent/nodes/extractor/node.py
Extraction Prompt src/agents/support_agent/nodes/extractor/prompt.py
Intent Router src/agents/support_agent/routes/intention/route.py
Routing Prompt src/agents/support_agent/routes/intention/prompt.py
Conversation Node src/agents/support_agent/nodes/conversation/node.py
Conversation Prompt src/agents/support_agent/nodes/conversation/prompt.py
Booking Agent src/agents/support_agent/nodes/booking/node.py
Booking Prompt src/agents/support_agent/nodes/booking/prompt.py
Appointment Tools src/agents/support_agent/nodes/booking/tools.py

Workflow Examples

Basic Agent src/agents/simple.py
Message Agent src/agents/simple-message.py
LLM Agent src/agents/simple-llm.py
RAG Agent src/agents/rag-basic.py
RAG Chaining src/agents/rag-basic-chaining.py
Structured RAG src/agents/rag-basic-structured-message.py
ReAct Agent src/agents/react.py
Orchestrator Workflow src/agents/orchestrator.py
Evaluator Optimizer src/agents/evaluator_optimizer_joke.py
Coder Checker src/agents/coder_checker.py

External Services

PostgreSQL
LLM Providers

Other components

Chat User

Connections 31

Chat User → Chat API: sends chat src/api/main.py
Chat API → Support Graph: invokes graph src/api/main.py
Chat User → Persistent Chat API: sends chat src/api/main-2.py
Persistent Chat API → Support Graph: builds and invokes src/api/main-2.py
Persistent Chat API → Checkpoint Lifecycle: requests checkpointer src/api/main-2.py
Checkpoint Lifecycle → PostgreSQL: persists checkpoints src/api/db.py
Support Graph → Customer Extractor: starts with src/agents/support_agent/support_agent.py
Support Graph → Intent Router: routes after extraction src/agents/support_agent/support_agent.py
Support Graph → Conversation Node: routes to src/agents/support_agent/support_agent.py
Support Graph → Booking Agent: routes to src/agents/support_agent/support_agent.py
Customer Extractor → Conversation State: reads and updates src/agents/support_agent/nodes/extractor/node.py
Conversation Node → Conversation State: reads messages src/agents/support_agent/nodes/conversation/node.py
Customer Extractor → Extraction Prompt: uses prompt src/agents/support_agent/nodes/extractor/node.py
Intent Router → Conversation State: reads messages src/agents/support_agent/routes/intention/route.py
Intent Router → Routing Prompt: uses prompt src/agents/support_agent/routes/intention/route.py
Conversation Node → Conversation Prompt: uses prompt src/agents/support_agent/nodes/conversation/node.py
Booking Agent → Appointment Tools: registers tools src/agents/support_agent/nodes/booking/node.py
Booking Agent → Booking Prompt: uses prompt src/agents/support_agent/nodes/booking/node.py
Customer Extractor → LLM Providers: calls model
Intent Router → LLM Providers: calls model
Conversation Node → LLM Providers: calls model
Basic Agent → LLM Providers: uses model
Message Agent → LLM Providers: uses model
LLM Agent → LLM Providers: uses model
RAG Agent → LLM Providers: uses model
RAG Chaining → LLM Providers: uses model
Structured RAG → LLM Providers: uses model
ReAct Agent → LLM Providers: uses model
Orchestrator Workflow → LLM Providers: uses model
Evaluator Optimizer → LLM Providers: uses model
Coder Checker → LLM Providers: uses model

Drawn by GitDiagram on November 12, 2026 from the repository's file tree, README and source files. Read as Markdown ·