Maestro — Your AI Video Master

An article, a complete video. Maestro automatically transforms articles/topics into short vertical videos with subtitles, with the entire process orchestrated by AI:

  • Script Writing — Claude breaks down the content into scenes (Scene-JSON)
  • Visuals — Seedance dynamic video clips + Seedream still frames (Ken Burns effect)
  • Voiceover — Fish TTS narration, supporting multiple languages including Chinese, English, Japanese, etc.
  • Music — Suno music bed, automatic ducking
  • Final Video — FFmpeg burns in subtitles + cross transitions + frame filling, outputting 9:16 / 16:9 / 1:1

A project can produce multiple language versions at once (reusing visuals, only adding voiceovers and rendering), suitable for content marketing, knowledge dissemination, and social media distribution.

How to Use

Asynchronous API: POST /maestro/videos submit → returns task_id → POST /maestro/tasks to poll for video (polling is free), or use callback_url for callback. See the API documentation for details.

Billing

Charged per completed video: ≤30s 18 points, >30s 32 points, +6 points for each additional language; task polling is free.