Skip to content
Conduit

mcp server

langchain-basics

intro to langchain

AnmolTomer/langchain-basicscommunitylistedunreviewed
Use tool

Set up

Runs on your computer. Pick your agent and operating system.

This project does not publish an install method we can turn into a config. Follow the README instead.

README

synced Oct 26, 2026

Foundation: Introduction to LangChain - Python

A comprehensive learning repository for building AI agents with LangChain, covering foundational concepts through production-ready multi-agent systems.

📚 Course Overview

This repository contains hands-on Jupyter notebooks and Python scripts for learning LangChain agent development. The course progresses from basic agent creation to advanced multi-agent orchestration with the Model Context Protocol (MCP).

Progress: 100% Complete (23/23 lessons)

🎯 Course Structure

✅ Module 1: Create Agent (Complete - 8/8)

Build foundational knowledge of LangChain agents, tools, and prompting techniques.

Topics Covered:

  • Lesson 1: Foundational Models - Agent invocation, prompting, and structured output
  • Lesson 2: Tools - Tool definition, integration with agents, and web search tools
  • Lesson 3: Short-Term Memory - Managing conversation context
  • Lesson 4: Multimodal Messages - Working with images, audio, and mixed media
  • Lesson 5: Personal Chef (Project) - Building a recipe recommendation agent

Key Files:

✅ Module 2: Advanced Agent (Complete - 8/8)

Learn sophisticated agent techniques including MCP, context management, and multi-agent systems.

Topics Covered:

  • Lesson 1: Model Context Protocol (MCP) - Standardized tool and data integration
  • Lesson 2: Context and State - Runtime context and stateful agents
  • Lesson 3: Multi-Agent Systems - Supervisor-subagent architecture
  • Lesson 4: Wedding Planner (Project) - Building a destination wedding planning agent team
  • Bonus: RAG Agent - Semantic search over PDF documents with vector embeddings
  • Bonus: SQL Agent - Natural language queries over a SQLite database

Key Concepts:

  • MCP as "USB-C for AI models" - universal protocol for tool integration
  • Context (immutable) vs State (mutable) management
  • Agent orchestration patterns
  • Multi-agent collaboration with LangGraph supervisor-subagent model
  • RAG (Retrieval Augmented Generation) with in-memory vector stores
  • Text-to-SQL with LangChain tool-calling agents

Key Files:

Documentation:

🚧 Module 3: Production-Ready Agent (In Progress - 7/7)

Advanced production concepts for deploying robust agent systems.

Topics Covered:

  • Lesson 1: Middleware - Production middleware patterns for agent customization
  • Lesson 2: Managing Messages - Long conversation management, summarization, and trimming
  • Lesson 3: Human-in-the-Loop - Approval checkpoints and human oversight patterns
  • Lesson 4: Dynamic Agents - Dynamic prompts, models, and tools with wrap-style middleware
  • Lesson 5: Email Assistant (Project) - Building an email inbox automation agent
  • Bonus: Agent Chat UI - Frontend app using LangGraph dev server

Key Concepts:

  • Middleware as customization hooks within agent execution loops
  • Node-style middleware (before/after hooks) vs wrap-style middleware (model-level)
  • Conversation summarization and message trimming for long-running agents
  • Human-in-the-loop patterns for approving sensitive actions
  • Dynamically swapping models, prompts, and tools at runtime

Key Files:

Documentation:

🚀 Getting Started

Prerequisites

  • Python 3.12 or 3.13
  • pip or uv package manager
  • API Keys (see Environment Setup)

Installation

  1. Clone the repository:

    git clone https://github.com/AnmolTomer/langchain-basics.git
cd langchain-basics

  2. Create virtual environment:

    # Using venv
python -m venv .venv
source .venv/bin/activate  # On Windows: .venv\Scripts\activate

# Or using uv (recommended)
uv venv
source .venv/bin/activate  # On Windows: .venv\Scripts\activate

  3. Install dependencies:

    # Using pip
pip install -r requirements.txt

# Or using uv (faster)
uv pip install -r requirements.txt

Environment Setup

  1. Copy the example environment file:

    cp example.env .env

  2. Add your API keys to .env:

    # Required
OPENAI_API_KEY='your_openai_api_key_here'
TAVILY_API_KEY='your_tavily_api_key_here'

# Optional (used in Module 1, Lesson 1)
ANTHROPIC_API_KEY='your_anthropic_api_key_here'
GOOGLE_API_KEY='your_google_api_key_here'

# Optional (for evaluation and tracing)
LANGSMITH_API_KEY='your_langsmith_api_key_here'
# LANGSMITH_TRACING=true
LANGSMITH_PROJECT=lca-lc-foundation

  3. Launch JupyterLab:

    jupyter lab

Windows Users: MCP Setup

If you're running MCP examples on Windows, you'll need to configure the event loop for subprocess support. See module-2/WINDOWS_MCP_FIX.md for detailed instructions.

Quick Fix:

import sys
import asyncio

if sys.platform == "win32":
    asyncio.set_event_loop_policy(asyncio.WindowsProactorEventLoopPolicy())
    # Then restart the kernel

🛠️ Key Technologies

  • LangChain - Framework for building LLM applications
  • LangGraph - Library for building stateful multi-agent applications
  • Model Context Protocol (MCP) - Anthropic's open protocol for tool integration
  • Tavily - AI-optimized search API
  • OpenAI / Anthropic / Google - LLM providers

📦 Project Structure

langchain-basics/
├── module-1/              # Foundational agent creation
│   ├── 1.1_foundation_models.ipynb
│   ├── 1.1_prompting.ipynb
│   ├── 1.2_tools.ipynb
│   ├── 1.3_memory.ipynb
│   ├── 1.4_Multimodal_Messages.ipynb
│   ├── 1.5_personal_chef.ipynb
│   ├── 1.5_personal_chef.py
│   └── README.md
├── module-2/              # Advanced agent techniques
│   ├── 2.1_mcp.ipynb
│   ├── 2.2_context_and_state.ipynb
│   ├── 2.3_multi_agents.ipynb
│   ├── 2.4_wedding_planners.ipynb
│   ├── bonus_rag.ipynb
│   ├── bonus_sql.ipynb
│   ├── resources/
│   │   ├── 2.1_mcp_server.py
│   │   ├── Chinook.db
│   │   └── acmecorp-employee-handbook.pdf
│   ├── Context_And_State_Documentation.md
│   ├── WINDOWS_MCP_FIX.md
│   └── README.md
├── module-3/              # Production-ready agent patterns
│   ├── 3.2_managing_messages.ipynb
│   ├── 3.3_human_in_the_loop.ipynb
│   ├── 3.4_dynamic_prompts.ipynb
│   ├── 3.4_dynamic_models.ipynb
│   ├── 3.4_dynamic_tools.ipynb
│   ├── 3.5_email_agent.ipynb
│   ├── 3.5_email_agent.py
│   ├── langgraph.json
│   ├── resources/
│   │   └── Chinook.db
│   └── README.md
├── env_utils.py           # Environment configuration utilities
├── pyproject.toml         # Project metadata and dependencies
├── requirements.txt       # Python package requirements
├── example.env            # Environment variables template
└── README.md             # This file

📖 Key Learning Resources

Module 1 Documentation

Module 2 Documentation

Module 3 Documentation

  • Module 3 README - Production-ready agent patterns and middleware concepts

External Resources

🎓 Learning Path

  1. Start with Module 1 - Work through notebooks 1.1 to 1.5 sequentially
  2. Complete the Personal Chef project - Apply foundational concepts
  3. Move to Module 2 - Learn MCP and advanced patterns
  4. Study Context & State - Read the comprehensive documentation
  5. Build Multi-Agent Systems - Implement supervisor-subagent patterns
  6. Complete the Wedding Planner - Multi-agent project
  7. Move to Module 3 - Learn production middleware and deployment patterns
  8. Master Dynamic Agents - Dynamic prompts, models, and tools
  9. Build the Email Assistant - Production-ready agent project

🤝 Contributing

This is a learning repository. Feel free to:

  • Report issues or bugs
  • Suggest improvements to documentation
  • Share your own agent implementations
  • Contribute fixes or enhancements

📝 Notes

  • Windows Users: Review WINDOWS_MCP_FIX.md before running MCP examples
  • API Costs: Be mindful of API usage when running notebooks
  • LangSmith: Optional but recommended for debugging and tracing agent behavior

🔧 Troubleshooting

Common Issues

Issue: NotImplementedError when running MCP on Windows

Issue: Import errors or missing modules

  • Solution: Ensure virtual environment is activated and all dependencies are installed

Issue: API key errors

  • Solution: Verify .env file exists and contains valid API keys

Issue: Jupyter kernel crashes

  • Solution: Restart kernel and re-run cells in order

📄 License

MIT License - see LICENSE file for details

Copyright (c) 2026 Anmol Tomer

🙏 Acknowledgments

This repository is based on the "Foundation: Introduction to LangChain - Python" course, covering modern agent development patterns with LangChain, LangGraph, and the Model Context Protocol.

Happy Learning! 🚀

For questions or discussions, please open an issue or refer to the module-specific README files.

↑↓ navigate openesc close