A sophisticated chatbot system built with LangGraph, featuring intent detection, personal information handling, and multi-agent workflow orchestration.

• 🤖 Multi-intent detection and handling
• 🔒 Personal information detection and secure storage
• 📝 Conversation history tracking
• 🔄 State management with MongoDB checkpointing
• 📊 Performance monitoring with Langfuse
• 🌐 Multi-language support

• Node.js (v14 or higher)
• Docker and Docker Compose
• MongoDB
• OpenAI API Key

1. Clone the repository:

git clone <repository-url>
cd <project-directory>

2. Install dependencies:

npm install

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

OPENAI_API_KEY=your_openai_api_key
MONGODB_URI=mongodb://root:example@localhost:27027
INTERNAL_API_URL=http://your-internal-api-url

4. Start the required services:

docker-compose up -d

The project consists of several key components:

• Intent Detection: Analyzes user input to determine the intended action
• Personal Info Detection: Identifies and securely handles personal information
• Multi-Agent System: Orchestrates different specialized agents for various tasks
• State Management: Maintains conversation state and history using MongoDB

To start the chatbot:

node semantic.js

The system supports various commands:

• Search number information
• Label search
• Contact counting
• Spam marking
• General chat interactions

The project includes several Docker services (reference to docker-compose.yml):

startLine: 1
endLine: 37

The project includes Langfuse integration for monitoring and analytics:

startLine: 9
endLine: 16

• LangGraph for the workflow orchestration framework
• OpenAI for the language models
• MongoDB for state management
• Langfuse for monitoring capabilities