🤖 LangChain + LangGraph AI Agent (TypeScript)

Build an AI-powered agent using LangChain and LangGraph in TypeScript. This project demonstrates how to set up an agent capable of performing reasoning, using tools, and navigating a graph-based workflow.

---

🛠 Features

• ⚙️ Agent setup with ChatPromptTemplate and agent_scratchpad
• 🧩 Custom tools (e.g., math operations)
• 🔄 LangGraph node/edge architecture
• 🌐 OpenAI (or other LLMs) powered reasoning
• ✨ Modular, clean TypeScript structure

---

🚀 Getting Started

1. Clone the Repository

git clone https://github.com/rehmat123/Langchain-typescript
cd langchain-typescript

2. Install Dependencies

pnpm install

3. Configure Environment

Create a .env file in the root:

OPENAI_API_KEY=your-openai-key

4. Run the Project

pnpm run langchain
pnpm run langraph

🧰 Tech Stack

• LangChain.js
  Framework for building context-aware, language model-driven applications.

• LangGraph
  A library for composing agents and chains as stateful graphs.

• TypeScript
  Strongly typed JavaScript for better developer experience and maintainability.

• OpenAI (or another LLM provider)
  Used as the underlying language model to power the agent's reasoning and responses.

✅ Future Ideas

• Add external API tools (e.g., weather, news)
• Integrate LangSmith for observability and debugging
• Docker support for easier deployment
• Web interface using React or Next.js

---

🙌 Contributing

Contributions are welcome!
If you find a bug, have a feature request, or want to add improvements, feel free to open an issue or submit a pull request.

---

📄 License

MIT License