Run TenderCheck AI with Docker Compose

​

Prerequisites

​

Docker Compose Configuration

services:
  backend:
    build:
      context: ./backend
      dockerfile: Dockerfile
      target: builder
    image: tendercheck-backend-dev
    ports:
      - "3000:3000"
    volumes:
      - ./backend:/app
      - /app/node_modules
    command: npm run dev
    environment:
      - NODE_ENV=development
      - PORT=3000

• build.context: ./backend — Docker uses backend/ as the build context, so the backend/Dockerfile is resolved relative to that directory.
• build.target: builder — Uses only the first (builder) stage of the multi-stage Dockerfile, which retains dev dependencies so that hot-reload tooling is available.
• ports: "3000:3000" — Maps container port 3000 to host port 3000.
• volumes: ./backend:/app — Bind-mounts the local backend/ directory into /app inside the container, enabling live code reloading without image rebuilds.
• /app/node_modules — An anonymous volume shadows the bind-mounted node_modules directory so the container’s installed packages are not overwritten by the host directory.
• command: npm run dev — Overrides the Dockerfile CMD and starts the backend in watch mode.

​

Running with Docker Compose

git clone https://github.com/elecodes/TenderCheck-AI.git
cd TenderCheck-AI

cp backend/.env.example backend/.env

TURSO_DB_URL=libsql://your-db-name.turso.io
TURSO_AUTH_TOKEN=your_turso_auth_token
GOOGLE_GENAI_API_KEY=your_gemini_api_key
GOOGLE_API_KEY=your_gemini_api_key
JWT_SECRET=your_strong_random_secret
ALLOWED_ORIGINS=http://localhost:3000

docker-compose up --build

docker-compose down

​

Individual Dockerfiles

​

Backend Dockerfile (backend/Dockerfile)

# Build Stage
FROM node:20-alpine AS builder

WORKDIR /app

COPY package*.json ./
RUN npm ci

COPY . .
RUN npm run build \
    && npm prune --production

# Production Stage
FROM node:20-alpine AS runner

WORKDIR /app

ENV NODE_ENV=production

COPY --from=builder --chown=node:node /app/package*.json ./
COPY --from=builder --chown=node:node /app/node_modules ./node_modules
COPY --from=builder --chown=node:node /app/dist ./dist

USER node

EXPOSE 3000

CMD ["node", "dist/index.js"]

• Node.js 20 on Alpine — Produces a small, security-hardened image.
• npm ci — Installs exact locked versions from package-lock.json for reproducible builds.
• npm run build + npm prune --production — Compiles TypeScript to dist/ then strips dev dependencies before the final image copy, minimizing image size.
• USER node — Runs the process as the unprivileged node user rather than root.
• EXPOSE 3000 — Documents the port; the actual binding is controlled by the PORT environment variable at runtime.

​

Root Dockerfile (Dockerfile)

​

Production Docker Deployment

• Set NODE_ENV=production in the backend container’s environment so that secure cookie flags and strict CORS enforcement are active.
• Set TURSO_DB_URL to your production Turso database URL. Use https:// if your runtime environment does not support WebSocket connections (see the warning below).
• Set ALLOWED_ORIGINS to your production frontend domain (e.g. https://yourdomain.com), without a trailing slash.
• Place a reverse proxy such as Nginx or Caddy in front of the containers to handle TLS termination. Expose only the frontend port publicly and proxy /api/* requests to the backend container internally.

yourdomain.com {
    reverse_proxy /api/* backend:3000
    reverse_proxy * frontend:80
}

​

Quick Start Script

#!/bin/bash
set -e

# Verify frontend build exists
if [ -d "frontend/dist" ]; then
  echo "Frontend build found at frontend/dist"
else
  echo "ERROR: frontend/dist NOT FOUND!"
fi

# Start Ollama in background
nohup ollama serve > /dev/null 2>&1 &
sleep 5

# Start backend
cd backend
export PORT=3000
exec npm start

docker-compose logs -f backend