A lightweight FTPS-based backend that automatically watches camera upload folders, converts the latest image from each camera to AVIF format, and stores them in a current/ directory for easy access (e.g., via Nginx).

Features

FTPS server for receiving image uploads
Automatic detection of the newest image per camera folder
Conversion to AVIF format using avifenc
Persistent storage of converted images in a shared current/ directory
Designed for integration with a frontend or Nginx for live webcam feeds

Getting Started

0) Import project files

Make sure the following files are present in your repository:

docker-compose.yml
Dockerfile
post_upload.sh

1) Configure user credentials

Edit docker-compose.yml and set your FTPS username and password:

environment:
  - BF_FTPS_VSFTPD_USER=username
  - BF_FTPS_VSFTPD_PASS=userpw

2) Make the post-upload script executable

chmod +x post_upload.sh

3) Build the Docker image

docker-compose build

4) Start the service

docker-compose up -d

Directory Structure

├── docker-compose.yml
├── Dockerfile
├── post_upload.sh
└── files/
    ├── msvcam1/      # uploads from camera 1
    ├── msvcam2/      # uploads from camera 2
    └── current/      # latest .avif images (msvcam1.avif, msvcam2.avif, ...)

How It Works

Cameras upload images via FTPS into files/msvcamN/ folders.
post_upload.sh periodically checks for the newest image in each folder.
If the newest image has changed (SHA256 hash comparison), it is converted to .avif format and saved to files/current/ as msvcamN.avif.
The current/ directory can be served directly by a web server (e.g., Nginx) for live display.

Example Nginx Integration

See MSV Webcam Frontend

Requirements

Docker
Docker Compose