LangGraphGo Examples

This directory contains 85+ examples demonstrating the features of LangGraphGo.

Table of Contents

• Basic Concepts
• State Management
• Graph Structure & Routing
• Parallel Execution
• Streaming & Events
• Persistence (Checkpointing)
• Human-in-the-Loop
• Pre-built Agents
• Programmatic Tool Calling (PTC)
• Memory
• RAG (Retrieval Augmented Generation)
• Tools & Integrations
• Advanced Patterns

---

Basic Concepts

• Basic Example - Simple graph with hardcoded steps
• Basic LLM - Integration with LLMs
• Arithmetic Example - Using LLM to calculate arithmetic expressions
• LangChain Integration - Using LangChain tools and models

State Management

• State Schema - Managing complex state updates with Schema and Reducers
• Custom Reducer - Defining custom state reducers for complex merge logic
• Smart Messages - Intelligent message merging with ID-based upserts
• Command API - Dynamic control flow and state updates from nodes
• Configuration - Using runtime configuration to pass metadata and settings

Graph Structure & Routing

• Conditional Routing - Dynamic routing based on state
• Conditional Edges - Using conditional edges for branching logic
• Subgraphs - Composing graphs within graphs
• Multiple Subgraphs - Managing multiple subgraph compositions
• Generic State Graph - Using generic types for type-safe state management
• Generic State Graph Listenable - Generic state graph with event listening
• Generic State Graph ReAct Agent - ReAct agent using generic types

Parallel Execution

• Parallel Execution - Fan-out/Fan-in execution with state merging
• Complex Parallel Execution - Advanced parallel execution with branches of varying lengths

Streaming & Events

• Streaming Modes - Advanced streaming with updates, values, and messages modes
• Streaming Pipeline - Building streaming data processing pipelines
• Listeners - Attaching event listeners to the graph
• Logger - Logging graph execution events
• LLM Streaming - Basic LLM streaming with callbacks
• LangChain Go Streaming - LangChain Go integration with streaming
• Adapter Streaming - Custom adapter streaming implementation

Persistence (Checkpointing)

• Memory Basic - In-memory checkpointing for state persistence
• Durable Execution - Crash recovery and resuming execution from checkpoints
• File Checkpointing - Checkpointing to file system
• File Checkpointing Resume - Resuming execution from file checkpoints

Human-in-the-Loop

• Human in the Loop - Workflow with interrupts and human approval steps
• Time Travel - Inspecting, modifying state history, and forking execution
• Dynamic Interrupt - Pausing execution from within a node using graph.Interrupt
• Payment Interrupt - Human approval workflow example for payment scenarios
• API Interrupt Demo - API-based interrupt demonstration for approval workflows

Pre-built Agents

Core Agent Patterns

• Create Agent - Easy way to create an agent with options
• Dynamic Skill Agent - Agent with dynamic skill discovery and selection
• ReAct Agent - Reason and Action agent using tools

Planning & Reasoning Agents

• Planning Agent - Intelligent agent that creates workflow plans based on user requests
• PEV Agent - Plan-Execute-Verify agent with self-correction and error recovery
• Reflection Agent - Iterative improvement agent that refines responses through self-reflection
• Tree of Thoughts - Search-based reasoning agent exploring multiple solution paths
• Mental Loop - Simulator-in-the-Loop agent that tests actions in a sandbox before execution
• Reflexive Metacognitive Agent - Self-aware agent with explicit self-model of capabilities
• Reflexive Metacognitive Agent CN - Chinese version of the reflexive metacognitive agent
• Manus Agent - Advanced agent for document processing workflows

Multi-Agent Systems

• Supervisor - Multi-agent orchestration using a supervisor pattern
• Swarm - Multi-agent collaboration using handoffs

Chat Agents

• Chat Agent - Multi-turn conversation agent with automatic session management
• Chat Agent Async - Asynchronous streaming chat agent with real-time LLM response streaming
• Chat Agent Dynamic Tools - Chat agent with runtime tool management

Programmatic Tool Calling (PTC)

Programmatic Tool Calling reduces latency by ~10x by having the LLM generate code instead of structured tool outputs.

• PTC Basic - Introduction to Programmatic Tool Calling
• PTC Simple - Simple PTC example with calculator and weather tools
• PTC Expense Analysis - Complex expense analysis scenario
• PTC + GoSkills - Integration of PTC with GoSkills for local tool execution

Memory

Conversation memory strategies for maintaining context across sessions.

• Memory Basic - Basic usage of LangChain memory adapters
• Memory Chatbot - Chatbot with LangChain memory integration
• Memory Strategies - Comprehensive guide to all 9 memory management strategies
• Memory Agent - Real-world agents using different memory strategies
• Memory Graph Integration - State-based memory integration in LangGraph workflows
• Memory - Additional memory examples

RAG (Retrieval Augmented Generation)

Basic RAG

• RAG Basic - Basic RAG implementation
• RAG Pipeline - Complete RAG pipeline
• RAG with LangChain - RAG using LangChain components

Advanced RAG

• RAG Advanced - Advanced RAG techniques
• RAG Conditional - Conditional RAG workflow
• RAG with Embeddings - RAG using embeddings
• RAG Query Rewrite - RAG with query rewriting for better retrieval
• LightRAG Simple - Basic LightRAG usage
• LightRAG Advanced - Advanced LightRAG patterns

RAG with Vector Stores

• RAG with VectorStores - RAG using LangChain VectorStores
• RAG with Chroma - RAG using Chroma database
• RAG with Chroma v2 - RAG using Chroma v2 database
• RAG with Chromem - RAG using chromem-go in-memory store
• RAG with Milvus - RAG using Milvus vector database
• RAG with Qwen Ranker - RAG with Qwen-based reranking
• RAG with Memory - RAG with conversation memory integration

GraphRAG (Knowledge Graph)

• RAG with FalkorDB Graph - RAG using FalkorDB knowledge graph with automatic entity extraction
• RAG with FalkorDB Simple - Simple RAG with FalkorDB using manual entity/relationship creation
• RAG with FalkorDB Fast - Optimized RAG with FalkorDB for fast queries
• RAG with FalkorDB Debug - Debug version with detailed logging
• RAG with FalkorDB Debug Query - Query debugging for FalkorDB RAG

Tools & Integrations

Search Tools

• Tavily Search - Using Tavily search tool with ReAct agent
• Exa Search - Using Exa search tool with ReAct agent
• Brave Search - Using Brave search API with agents

Other Integrations

• GoSkills Integration - Integrating GoSkills as tools for agents
• MCP Agent - Using Model Context Protocol (MCP) tools with agents
• Context Store - Managing context with external stores

Advanced Patterns

• Visualization - Generating Mermaid diagrams for graphs

---

Running Examples

Most examples can be run directly:

cd examples/<example_name>
go run main.go

Some examples may require additional setup:

1. API Keys: Search tools (Tavily, Exa, Brave) require API keys
2. Databases: FalkorDB examples require FalkorDB to be running
3. Vector Stores: Chroma examples require Chroma DB to be running
4. Redis: Some checkpointing examples require Redis

Example Categories

Category	Count	Description
Basic Concepts	3	Getting started examples
State Management	5	State schema and reducers
Graph Structure	7	Routing, subgraphs, generics
Parallel Execution	2	Concurrent execution patterns
Streaming & Events	7	Real-time data flow
Persistence	4	Checkpointing and recovery
Human-in-the-Loop	5	Interactive workflows
Pre-built Agents	18	Ready-to-use agent patterns
PTC	4	Programmatic tool calling
Memory	6	Conversation memory strategies
RAG	22	Retrieval-augmented generation
Tools & Integrations	6	External service integrations
Advanced	1	Visualization and debugging
Total	85+	Comprehensive examples