LangGraph Models Repository

Welcome to the LangGraph Models repository by vikas-kashyap97.
This repository presents modular implementations for building complex LLM workflows using LangGraph—an extension of LangChain for stateful, agent-based, and memory-aware reasoning.

---

📂 Repository Structure

This project contains step-by-step LangGraph-based LLM applications, categorized by feature complexity and real-world use cases.

---

📦 Module Overview

1_Introduction/

Basic example of an LLM-powered ReAct agent using LangGraph.

• react_agent_basic.py: Minimal ReAct implementation for understanding core LangGraph concepts.

---

2_basic_reflection_system/

A simple agent loop with self-reflection capabilities.

• basic.py: Defines a reflection loop logic.
• chains.py: Defines LangChain chains used within the loop.

---

3_structured_outputs/

Demonstrates how to work with structured outputs from LLMs using Pydantic.

• structured_outputs.py: LLM output parsing and validation using Pydantic.
• pydantic_outputs.json: Sample output for schema reference.

---

4_reflexion_agent_system/

More advanced reflexion loop with schema-enforced tools and agent planning.

• chains.py, schema.py: Tool and output definitions.
• reflexion_graph.py: Full graph implementation.
• execute_tools.py: Tool invocations and execution layer.

---

5_state_deepdive/

Hands-on with LangGraph state management.

• 1_basic_state.py: Introduction to LangGraph state handling.
• 2_complex_state.py: Managing multiple variables and transitions across nodes.

---

6_react_agent/

Advanced ReAct agent architecture with execution planning and modular reasoning.

• agent_reason_runnable.py: Defines reasoning logic as reusable runnables.
• nodes.py, react_graph.py, react_state.py: Complete node-graph pipeline.

---

7_chatbot/

Multiple chatbot implementations with varying checkpoint and memory strategies.

• 1_basic_chatbot.py: Stateless chatbot using LangGraph.
• 2_chatbot_with_tools.py: Tool-augmented chatbot.
• 3_chat_with_in_memory_checkpointer.py: Memory persistence using in-memory checkpoints.
• 4_chat_with_sqlite_checkpointer.py: Chat persistence with SQLite.
• checkpoint.sqlite: Database for chat state saving.

---

8_human-in-the-loop/

Integrating human feedback in LangGraph loops.

• 1_using_input(): Capturing manual input mid-process.
• 2_command.py, 3_resume.py: Command flow controls.
• 5_multiturn_conversation.py: Rich multi-turn human-AI dialogue.

---

9_RAG_agent/

Retrieval-Augmented Generation (RAG) agents with progressive complexity.

• 1_basic.py: Basic RAG agent flow.
• 2_classification_driven_agent.py: RAG agent using classification to guide actions.
• 3_rag_powered_tool_calling.py: Tool-augmented RAG agent using knowledge retrieval.
• 4_advanced_multi_step_reasoning.py: Complex reasoning over retrieved knowledge.

---

10_multi_agent_architecture/

Collaborative multi-agent systems for task delegation and supervision.

• 1_subgraphs.py: Modular subgraph-based agent interactions.
• 2_supervisor_multiagent_workflow.py: Supervisor-worker agent architecture.

---

11_streaming/

Streaming outputs from LLM agents in real-time.

• 1_stream_events.py: Event-driven stream interface for agent outputs.

🔧 Environment Setup

# Create and activate a virtual environment
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

---

📚 Official Resources

• LangGraph Documentation
• LangChain GitHub
• LangGraph GitHub
• Official LangChain Examples

---

💡 Author

Developed with ❤️ by vikas-kashyap97
Feel free to contribute or raise issues to enhance this LangGraph resource hub.