BrandKit

Appearance

Running with Docker ​

Docker is the supported deployment path. The image bundles OpenCV and Pillow system libraries required for processing.

Quick start ​

bash
git clone https://github.com/fabriziosalmi/brandkit.git
cd brandkit
docker compose up -d --build

BrandKit is available on http://localhost:8000.

bash
docker compose logs -f brandkit   # follow logs
docker compose down               # stop
docker compose down -v            # stop and remove volumes

Docker Compose configuration ​

yaml
services:
  brandkit:
    build: .
    container_name: brandkit
    ports:
      - "8000:8000"
    volumes:
      - ./static/uploads:/app/static/uploads
    environment:
      - FLASK_ENV=production        # legacy variable
    restart: unless-stopped

Key configuration details:

  • Upload volume: ./static/uploads on the host is the container upload directory. Generated assets are stored directly in this directory.
  • Port binding: The internal container port is 8000. To bind specifically to host loopback, specify "127.0.0.1:8000:8000".
  • Cleanup handling: Background cleanup runs under Gunicorn and is configured via BRANDKIT_CLEANUP_* environment variables.

Running without Compose ​

bash
docker build -t brandkit .

docker run -d \
  --name brandkit \
  -p 127.0.0.1:8000:8000 \
  -v "$(pwd)/static/uploads:/app/static/uploads" \
  -v "$HOME/.u2net:/root/.u2net" \
  -e BRANDKIT_MAX_UPLOAD_MB=32 \
  --restart unless-stopped \
  brandkit

Mounting $HOME/.u2net persists downloaded ONNX models across container rebuilds.

Image specification ​

The Dockerfile builds on python:3.11-slim and installs system dependencies for rembg and OpenCV (libgl1, libglib2.0-0, libjpeg-dev, zlib1g-dev, libpng-dev, libwebp-dev, and OpenCV development headers). It then installs rembg, onnxruntime, opencv-python-headless, numpy, and dependencies listed in requirements.txt.

Initial builds take several minutes, resulting in an image size of approximately 2 GB due to the ONNX Runtime and scientific Python stack.

Entrypoint process ​

The container executes entrypoint.sh, selecting a WSGI runner in priority order:

  1. Gunicorn (when available on PATH): gunicorn --bind 0.0.0.0:$PORT --worker-tmp-dir /dev/shm app:app
  2. Flask CLI (when FLASK_APP is set): flask run
  3. Python runtime: imports app.py and initiates app.run(host="0.0.0.0")

The entrypoint creates static/uploads if absent and respects the PORT environment variable (default: 8000).

Configure secret key before scaling workers

The default configuration runs 1 Gunicorn worker. Before increasing worker count, configure BRANDKIT_SECRET_KEY to ensure consistent CSRF session verification across workers.

Updating ​

bash
cd brandkit
git pull
docker compose up -d --build

Upload directories in the bind mount persist across rebuilds.

Health check ​

The /format-info endpoint serves as a lightweight HTTP GET probe without requiring CSRF authentication:

yaml
    healthcheck:
      test: ["CMD", "python", "-c",
             "import urllib.request,sys; sys.exit(0 if urllib.request.urlopen('http://127.0.0.1:8000/format-info').status==200 else 1)"]
      interval: 30s
      timeout: 5s
      retries: 3
      start_period: 40s

Resource limits ​

Background removal on high-resolution images can allocate significant RAM. On resource-constrained hosts, specify explicit limits:

yaml
    deploy:
      resources:
        limits:
          memory: 4G

Configure BRANDKIT_MAX_UPLOAD_MB accordingly. See Environment variables.