LangChain 1.0 Guide & RAG Demo

A comprehensive guide and demonstration of LangChain 1.0 features, including the new create_agent workflow, Human-in-the-Loop (HITL) middleware, and Multi-Agent systems.

This project also includes a RAG pipeline that ingests the official LangChain 1.0 announcement blog post to answer questions about the framework updates.

Features Covered

• Standard Agents: Using create_agent with custom tools.
• Human-in-the-Loop: Pausing execution for approval using HumanInTheLoopMiddleware.
• RAG: Integrating Qdrant and OpenAIEmbeddings to query external documents.
• Multi-Agent Orchestration: Composing agents (Supervisor -> Researcher + Writer).

Pre-Reqs

You'll need to ensure you have uv installed before proceeding.

# Install dependencies
uv sync

You will also need an OpenAI API Key (OPENAI_API_KEY).

Running the Example

The core logic is in the src/ directory, but the best way to explore is via the Jupyter Notebook.

1. Launch the Notebook:

   uv run jupyter notebook langchain_v1_0_guide.ipynb

2. Run the Cells:

   • The notebook will guide you through setting up your environment.
   • It will demonstrate a simple agent, then a HITL agent, then a RAG agent, and finally a Multi-Agent system.

Project Structure

• src/agent.py: Contains the agent construction logic (create_agent, middleware setup).
• src/rag.py: Handles the RAG pipeline (loading the blog post, chunking, vector store).
• src/tools.py: Simple example tools (get_weather, magic_calculator).
• langchain_v1_0_guide.ipynb: The interactive guide.
• pyproject.toml: Dependency management.
• langgraph.json: Configuration for deploying through LangSmith.

Deploy with LangSmith

LangSmith is the fastest way to turn agents into production systems.

1. Prerequisites

• A GitHub account
• A LangSmith account

2. Create a Repository

Your application's code must reside in a GitHub repository. Push this code to a new repository.

3. Deploy

1. Go to LangSmith and navigate to the Deployments tab.
2. Click + New Deployment.
3. Connect your GitHub repository.
4. LangSmith will detect the langgraph.json file and automatically configure your deployment.
   • Entrypoint: langgraph.json defines multiple graphs (agent, hitl_agent, rag_agent, multi_agent). You can choose which one to expose or deploy them all.
5. Set your environment variables (e.g., OPENAI_API_KEY) in the deployment settings.
6. Click Deploy.

Once deployed, you can interact with your agent via the LangSmith Studio, API, or SDK.

Credits

Built using LangChain and LangGraph.