CLI reference

Drive the whole Podcli pipeline from your terminal. One command transcribes, picks moments, crops to the speaker, and burns captions in.

  • 01Guides
  • 024 sections
  • 03Markdown reference

Everything the studio does is available from the terminal. The one command you need most:

That transcribes, picks the best moments, crops to the face, burns captions in, and writes upload-ready clips. Nothing leaves your machine.

Common flags#

Use an existing transcript instead of transcribing, and cap the clip count:

Set caption style, crop mode, and output format:

Force thumbnails on or off:

FlagWhat it does
--transcript <file>Use a .txt / .srt / .vtt transcript instead of transcribing.
--top <n>Cap the number of clips.
--caption-style <style>branded, hormozi, karaoke, or subtle. See captions and formats.
--crop <mode>center, face, or speaker.
--format <ratio>vertical (9:16), horizontal (16:9), or square (1:1).
--logo <asset or file>Overlay a logo on every clip. Accepts a registered asset name or a path.
--intro <asset or file>Prepend an intro video before every clip.
--outro <asset or file>Append an outro video after every clip.
--thumbnails / --no-thumbnailsForce thumbnail generation on or off.

Logo, intro, and outro all accept a registered asset name (see Assets) or a direct file path. Leave them off to use whatever you've marked as the default asset.

Presets#

Save a set of render settings once and reuse it:

Assets#

Register logos, outros, intros, and music once, then reference them by name anywhere. The same library backs the studio Assets page and the manage_assets MCP tool.

Then use a name instead of a path:

Other actions: podcli assets rename <old> <new>, podcli assets undefault <name>, and podcli assets remove <name>.

Other commands#

podcli opens the interactive menu with Open Web UI. podcli ui opens the studio directly at localhost:3847.

podcli uninstall removes all managed data by default. Run it with --dry-run first to preview exactly what will be removed.