AI Agent with LangGraph.js and MongoDB

An advanced AI agent implementation using LangGraph.js and MongoDB Atlas Vector Search for intelligent HR assistance and employee data management.

🌟 Features

• AI-powered conversational agent using LangGraph.js
• MongoDB Atlas Vector Search for semantic employee search
• Persistent conversation state using MongoDB checkpoints
• Employee data management with vector embeddings
• RESTful API endpoints for chat interactions
• TypeScript implementation for type safety
• Extensible tool system for agent capabilities

📋 Prerequisites

• Node.js >= 14
• MongoDB Atlas account
• OpenAI API key
• Anthropic API key
• npm or yarn package manager

🛠️ Tech Stack

• LangGraph.js - AI agent framework
• MongoDB Atlas - Database and vector search
• Express.js - API server
• TypeScript - Programming language
• Anthropic Claude - Language model
• OpenAI - Embeddings generation

🚀 Getting Started

1. Clone the repository:

git clone <repository-url>
cd ai-agent-project

2. Install dependencies:

npm install

3. Create a .env file in the project root:

OPENAI_API_KEY=your-openai-api-key
ANTHROPIC_API_KEY=your-anthropic-api-key
MONGODB_ATLAS_URI=your-mongodb-atlas-connection-string

4. Set up MongoDB Atlas Vector Search index:

• Create a vector index named "vector_index"
• Use the following index definition:

{
  "fields": [
    {
      "numDimensions": 1536,
      "path": "embedding",
      "similarity": "cosine",
      "type": "vector"
    }
  ]
}

5. Seed the database:

npx ts-node seed-database.ts

6. Start the server:

npx ts-node index.ts

📁 Project Structure

├── .env                    # Environment variables
├── index.ts               # Express server setup
├── agent.ts               # AI agent implementation
├── seed-database.ts       # Database seeding script
├── package.json           # Project dependencies
└── tsconfig.json          # TypeScript configuration

🔌 API Endpoints

Start New Chat

POST /chat
Content-Type: application/json

{
  "message": "Your initial message"
}

Continue Conversation

POST /chat/:threadId
Content-Type: application/json

{
  "message": "Your follow-up message"
}

💻 Usage Example

# Start a new conversation
curl -X POST \
  -H "Content-Type: application/json" \
  -d '{"message": "Build a team to make an iOS app, and tell me the talent gaps."}' \
  http://localhost:3000/chat

# Continue conversation
curl -X POST \
  -H "Content-Type: application/json" \
  -d '{"message": "What are the talent gaps?"}' \
  http://localhost:3000/chat/<threadId>

🛠️ Development

Adding New Tools

1. Create a new tool in agent.ts:

const newTool = tool(
  async ({ param1, param2 }) => {
    // Tool implementation
    return result;
  },
  {
    name: "tool_name",
    description: "Tool description",
    schema: z.object({
      param1: z.string(),
      param2: z.number()
    })
  }
);

2. Add the tool to the tools array:

const tools = [employeeLookupTool, newTool];

Modifying Agent Behavior

Update the prompt template in callModel function:

const prompt = ChatPromptTemplate.fromMessages([
  [
    "system",
    `Your updated system message here...`
  ],
  new MessagesPlaceholder("messages")
]);

🔒 Security Considerations

• Secure API keys in environment variables
• Implement rate limiting for API endpoints
• Validate user input
• Sanitize tool outputs
• Use secure MongoDB connection string
• Implement proper error handling

⚙️ Configuration

Environment Variables

• OPENAI_API_KEY: OpenAI API key for embeddings
• ANTHROPIC_API_KEY: Anthropic API key for Claude LLM
• MONGODB_ATLAS_URI: MongoDB Atlas connection string
• PORT: Server port (default: 3000)

MongoDB Collections

• employees: Stores employee data and embeddings
• checkpoints: Stores conversation states

🤝 Contributing

1. Fork the repository
2. Create your feature branch (git checkout -b feature/amazing-feature)
3. Commit your changes (git commit -m 'Add amazing feature')
4. Push to the branch (git push origin feature/amazing-feature)
5. Open a Pull Request

📝 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

• Built with LangGraph.js
• MongoDB Atlas Vector Search
• Anthropic Claude API
• OpenAI API

📧 Support

For questions and support, please open an issue in the repository or contact the maintainers.