• Transcode media using ffmpeg
• Optionally use hardware acceleration
• Dynamic queue management
• Progress monitor
• Built with Flask/React

• This assumes that you already know the ffmpeg command you want to run on your collection. Maybe you want to reduce the CRF value, reduce bitrate or convert to x265.
• This is not meant to be run on all your media files, pick and choose what makes sense. Depending on your input file and your ffmpeg configuration, your ffmpeg command may or may not compress your files or will even have a degrading effect.
• This will default to .mkv format and will overwrite the original files you add to the queue. So when in doubt, test your configuration and run the transcode on a copy and not on the original file.

This is a docker container. Configure your volumes and environment variables as described below, there are compose examples further down. Help with adding additional instructions.

• /data: Application data volume to store the queue database
• /media: Root path for your media collection folder
• /cache: Temporary path to store the file while transcoding, should be on the same drive as /media.

• TT_FFMPEG: JSON formatted Key Value config pairs to be passed to ffmpeg.
• TT_HWACCEL (optional): Define method for hardware acceleration, e.g. cuda
• TT_PORT (optional): Overwrite the default port 8000.

Resources to enable GPU support for a docker container, adjust to your environment:

• Docker docs
• Arch Wiki

Contribute to add additional hardware configurations.

version: '3.5'

services:
  tiny-transcoder:
    container_name: tiny-transcoder
    restart: unless-stopped
    image: bbilly1/tiny-transcoder
    volumes:
      - /media/docker/tiny-transcoder:/data
      - /media/mount/cache:/cache
      - /media/mount:/media
    ports:
      - 127.0.0.1:8000:8000
    environment:
      TT_FFMPEG: '{"c:v": "libx265", "crf": 23, "preset": "slow", "vf": "scale=-1:1080"}'

version: '3.5'

services:
  tiny-transcoder:
    container_name: tiny-transcoder
    restart: unless-stopped
    image: bbilly1/tiny-transcoder
    volumes:
      - /media/docker/tiny-transcoder:/data
      - /media/mount/cache:/cache
      - /media/mount:/media
    ports:
      - 127.0.0.1:8000:8000
    environment:
      NVIDIA_VISIBLE_DEVICES: all
      NVIDIA_DRIVER_CAPABILITIES: gpu,video
      TT_HWACCEL: cuda
      TT_FFMPEG: '{"c:v": "hevc_nvenc", "crf": 23, "preset": "slow", "vf": "scale=-1:1080"}'
    deploy:
      resources:
   	 reservations:
 		 devices:
 		 - driver: nvidia
   		 capabilities:
     		 - gpu
     		 - video

The queue is meant to be transient, clear the queue regularly, keep your queue to a manageable size. There probably won't be a migration path for your database. Check the release notes if you have to clear your queue.db.

Click on the add icon to open a form to add media files to your queue. Click on the add icon again, to hide the form.

Absolute/relative: If you add an absolute path, make sure to prefix that with /media, as the path is absolute from inside the container. If you add a relative path that will be relative from the /media folder.

Folder/file: You can add a folder to the queue to add all files of the folder. Make sure there are only media files in the folder. That is not recursive. Or add file by file.

If you have any files in the queue, a start icon will show to start the queue. That will run the queue until there are no more pending items.

Click on the delete icon to show options to delete the queue. You can delete by state pending, paused, completed, or delete all. Click on the delete icon again to hide the options.

This is very much inspired by the superior Tdarr project and is a less complex, less feature rich, tiny project to transcode your media.