CantusManualBack to the site

Exporting a dancing video

Back to the manual · Leer en español · See also: the dancer · the stage · importing characters

File ▸ Export video… makes a video of your dancers with the song: a vertical clip for TikTok, Reels or Shorts, a square one, a wide one for YouTube, or a transparent layer for a video editor. What you see in the preview is what the file will hold, frame for frame — the dancers move exactly as they do on the screen, their automation included.

The window

The footer says how long the export should take and how big the file will be; during the export, how far it is and how long is left. Stop stops it and leaves nothing behind.

In sync, to the sample

Each frame is drawn at its exact time and the sound starts on the same sample: a beat's step lands on its beat in every player and on every platform. One exception you might notice when comparing very closely: Windows' own player (Media Player, Films & TV) plays the sound of any AAC MP4 about 21 ms late — files from other programs too — because it ignores the instruction that trims the encoder's start. What you upload is exact.

Where the encoders come from

Cantus ships no video encoder of its own (H.264 and AAC are licensed per copy):

For agents

An assistant connected through MCP can export a video with export_video (format, shape, frame rate, quality, background, title, range, file) and follow it with get_video_export. Like every tool that writes files, it is refused until you allow files in Settings ▸ AI, and you are always asked.