Quick Start

Get the TFG-Chatbot running in minutes.

This is the fastest way to get all services running:

# 1. Start all services
docker compose up -d

# 2. Wait for services to be healthy (check logs)
docker compose logs -f

# 3. Initialize Ollama with embeddings model (first time only)
./scripts/init_ollama.sh

# 4. Seed demo users (optional)
python scripts/seed_users.py

Service URLs

Service URL Description
Frontend http://localhost:3000 Web interface
Backend API http://localhost:8000/docs Swagger documentation
Chatbot API http://localhost:8080/docs Internal chatbot API
RAG Service http://localhost:8081/docs RAG operations API

Demo Credentials

User Password Role
admin admin123 Admin
profesor admin123 Professor (subject: ISE)
estudiante admin123 Student (subject: ISE)

Option B: Local Development (Without Containers)

For development with hot-reloading:

Start Infrastructure Services 

# Start MongoDB, Qdrant, and Ollama containers only
docker compose up -d mongo qdrant ollama

Start Backend Services

Open separate terminals for each service:

Terminal 1: Backend Gateway

cd backend
uv run uvicorn backend.api:app --reload --port 8000

Terminal 2: Chatbot Service

cd chatbot
uv run uvicorn chatbot.api:app --reload --port 8080

Terminal 3: RAG Service

cd rag_service
uv run uvicorn rag_service.api:app --reload --port 8081

Terminal 4: Frontend

cd frontend
npm run dev

Option C: Quick Test with Script

For quick testing of the backend only:

./scripts/run_fastAPI.sh

This starts the backend gateway on port 8080 with hot-reload.

Verify Services

Check Health Endpoints 

# Backend
curl http://localhost:8000/health

# Chatbot
curl http://localhost:8080/health

# RAG Service
curl http://localhost:8081/health

Check Container Status 

docker compose ps

Expected output:

NAME             STATUS          PORTS
tfg-gateway      Up (healthy)    0.0.0.0:8000->8000/tcp
tfg-chatbot      Up (healthy)    0.0.0.0:8080->8080/tcp
tfg-rag-service      Up (healthy)    0.0.0.0:8081->8081/tcp
frontend         Up              0.0.0.0:3000->80/tcp
mongo-service    Up              0.0.0.0:27017->27017/tcp
qdrant-service   Up              0.0.0.0:6333->6333/tcp
ollama-service   Up              0.0.0.0:11434->11434/tcp

First Steps After Setup

1. Access the Frontend

Open http://localhost:3000 and log in with demo credentials.

2. Try the Chat

Send a message to the chatbot. Try questions like:

  • “¿Qué es Docker?”
  • “Explica qué es un contenedor”
  • “¿Cuál es la diferencia entre Docker y una máquina virtual?”

3. Explore the API

Visit http://localhost:8000/docs to see all available endpoints.

4. Upload Documents (Optional)

Use the RAG Service API to upload documents for semantic search:

curl -X POST http://localhost:8081/index \
  -H "Content-Type: multipart/form-data" \
  -F "file=@document.pdf" \
  -F "subject=ISE"

Stopping Services 

# Stop all containers
docker compose down

# Stop and remove volumes (clean reset)
docker compose down -v

Next Steps