← All skills & workflows

video

avatar-video

Generates lip-synced talking head videos with MuseTalk (local) or HeyGen (cloud), with optional Remotion compositing.

Generate lip-synced talking head videos through a backend router — MuseTalk (local GPU-based, default) or HeyGen (cloud-hosted, no GPU needed). Includes video dubbing into 175+ languages via HeyGen's video-translate.

Difficulty
expert
Input
conversation
Output
video-script
Needs
api-key, install
Used by
Head of Marketing, Content Manager

What it needs

  • source (Simple avatar) — Script or message text
  • backend — Backend choice: musetalk (default, requires GPU) | heygen (cloud, requires HEYGEN_API_KEY)
  • avatar-id — HeyGen avatar ID (if using heygen backend and reusing existing avatar)
  • source (Narrated slides) — Slides JSON + narration pairs
  • source (Composited video) — Script text (will run generate route first, then composite)
  • video-path — Path to or URL of video to translate
  • target-languages — Target language(s): e.g. ['es', 'fr', 'ja'] — full list via `heygen video-translate languages list`

What you get

  • avatar-video (video)
  • composited-video (video)
  • dubbed-video (video)
  • dubbing-log (json)
  • narrated-video (video)

Ask it like this

  • Create a talking head video from this photo and script
  • Generate a lip-synced avatar for our product demo
  • Make a spokesperson video using MuseTalk from this image