SmartCustomerService is a Retrieval-Augmented Generation (RAG) powered GenAI application designed to enhance customer interactions by providing a platform where users can inquire about company-specific information. Administrators can securely upload and manage company data, while users can select a company from a list and engage in informative conversations based on the provided details. Users can interact directly with a base Large Language Model (LLM) DeepSeek if no company information is available.
SmartCustomerService-demo.mp4
- Admin Access: Authorized administrators can log in to securely upload and manage company information.
- User Interaction: Users can view a list of available companies and engage in conversations to inquire about specific company details.
- Fallback Chat: In the absence of company-specific information, users can interact directly with a base LLM powered by DeepSeek.
- Backend: FastAPI
- Frontend: React with Next.js
- LLM and Embedding Models: DeepSeek-v2 and bge-large deployed locally using Ollama
- Python 3.10 or higher
- Node.js 14 or higher
- npm or yarn package manager
- Ollama installed for local deployment of LLM and embedding models
-
Clone the repository:
git clone https://github.com/medxiaorudan/SmartCustomerService.git cd SmartCustomerService -
Ollama Setup for Local Deployment:
-
Install Ollama:
Download and install Ollama onto your platform. Instructions are available on the Ollama website.
-
Deploy DeepSeek-v2 Model Locally:
Use Ollama to pull and run the DeepSeek-v2 model:
ollama pull deepseek-v2
This command downloads the DeepSeek-v2 model for local use.
-
Deploy bge-large Embedding Model Locally:
Similarly, pull the bge-large embedding model:
ollama pull bge-large
This command downloads the bge-large embedding model for local use.
-
-
Backend Setup:
-
Create and activate a virtual environment:
python -m venv env source env/bin/activate # On Windows, use 'env\Scripts\activate'
-
Create a
.envfile in the repo root:cp .env.example .env
Set
ADMIN_PASSWORDandJWT_SECRETto real values before starting the API. KeepEMB_CONFIG_LISTpointed athttp://localhost:11434andCONFIG_LISTpointed athttp://localhost:11434/v1. -
Install the required dependencies:
pip install -r requirements.txt
-
Start the FastAPI server:
uvicorn main:app --host 0.0.0.0 --port 8000 --reload
The backend API will be accessible at
http://127.0.0.1:8000. -
Verify the backend before opening the frontend:
curl http://127.0.0.1:8000/health
You should get a JSON response with
activeSessionsandlastActivityTs.
-
-
Frontend Setup:
-
Navigate to the frontend directory:
cd smart-customer-service -
Install the required dependencies:
npm install
-
Start the Next.js development server:
npm run dev
The frontend will be accessible at
http://localhost:3000.
-
- Admin Access: Navigate to the admin login page at
http://localhost:3000/adminand enter your credentials to upload or manage company information. - User ChatBox: Visit the homepage at
http://localhost:3000/user, select a company from the list, and start asking questions related to that company. If no company information is available, you can interact directly with the base LLM powered by DeepSeek.
For production deployment, consider using Docker to containerize both the backend and frontend services. Ensure that environment variables are properly set and that the services are configured to run in a production environment.
Contributions are welcome! Please fork the repository and create a pull request with your proposed changes. Ensure that your code adheres to the project's coding standards and includes appropriate tests.
This project is licensed under the MIT License. See the LICENSE file for more details.
For more information, visit the project repository at https://github.com/medxiaorudan/SmartCustomerService.
- Admin: Fix file upload
- Docker: Setup Docker for hosting of the backend
- Vite: Look into hosting frontend