Skip to main content

Quick start

This walkthrough saves one public YouTube video together with the context needed to identify it later.

1. Check the tools​

Open PowerShell, Terminal, or your preferred shell:

yt-dlp --version
ffmpeg -version
deno --version

If any command is missing, follow Prerequisites.

2. Choose an archive folder​

Create a folder on a drive with enough free space, then open it in your terminal.

On Linux or macOS:

mkdir -p "Suisei Archive"
cd "Suisei Archive"

On PowerShell:

New-Item -ItemType Directory -Path "Suisei Archive"
Set-Location "Suisei Archive"

3. Download media and context​

Replace VIDEO_ID with the ID from the YouTube URL:

yt-dlp \
  --write-info-json \
  --write-description \
  --write-thumbnail \
  --write-subs \
  --write-auto-subs \
  --sub-langs "en.*,ja.*,id.*" \
  -o "%(upload_date)s - %(title)s [%(id)s].%(ext)s" \
  "https://www.youtube.com/watch?v=VIDEO_ID"

PowerShell accepts the command on one line. If you split it across lines, replace each trailing \ with a PowerShell backtick.

yt-dlp may create several files with the same base name: the video, metadata (.info.json), description, thumbnail, and English, Japanese, or Indonesian subtitle tracks. That is intentional; adjust --sub-langs if you need other languages.

4. Verify the result​

  • Play the media from beginning to end, or at least seek through its start, middle, and end.
  • Confirm that it has both picture and sound.
  • Open the description and check that the info JSON is not empty.
  • Keep every sidecar beside the media file.

Once this works, read the full yt-dlp guide, then add checksums and backups. For a stream that has not begun, use the scheduled live workflow instead.