AGI Chatbot (React + Node.js + FastAPI)

This project uses a 3-service architecture:

• apps/web-react: React frontend (Vite)
• apps/api-node: Node.js API gateway (Express)
• apps/ml-fastapi: FastAPI NLP inference service (PyTorch + NLTK)

Architecture

1. Frontend sends chat message to POST /api/chat on Node.
2. Node forwards the request to FastAPI POST /chat.
3. FastAPI runs model inference and returns response.

Project Structure

apps/
  web-react/
  api-node/
  ml-fastapi/

docker-compose.yml
README.md

Prerequisites

• Node.js 20+
• Python 3.12+ (or compatible with your environment)
• npm
• Optional: Docker Desktop

Run Locally (without Docker)

Open 3 terminals.

1. Start FastAPI ML service:

cd apps/ml-fastapi
python -m venv .venv
.\.venv\Scripts\Activate.ps1
pip install -r requirements.txt
uvicorn app.main:app --reload --port 8000

2. Start Node API service:

cd apps/api-node
npm install
copy .env.example .env
npm run dev

3. Start React frontend:

cd apps/web-react
npm install
copy .env.example .env
npm run dev

Open http://localhost:5173.

Run with Docker

docker compose up --build

Then open http://localhost:5173.

Health Checks

• FastAPI: GET http://localhost:8000/health
• Node API: GET http://localhost:4000/health

Main Endpoints

• Node chat endpoint: POST http://localhost:4000/api/chat
• FastAPI chat endpoint: POST http://localhost:8000/chat

Notes

• NLP model assets are stored in apps/ml-fastapi/assets/.
• Environment templates are provided as .env.example in each app where needed.