Use case: a 3-minute 1920s jazz music video made fully locally on one RTX 4070
Creator @Jeuner made a 3-minute, 23-shot music video for Bibi Beyond, an invented 1920s jazz singer, entirely on one 12 GB RTX 4070. MiniMax H3 animated each shot from a keyframe and the song, including the lip-sync, and Claude Code ran the direction, editing and pipeline.
What it does
The stack: Krea2 Turbo for the character and the club, Qwen Image Edit 2509 for keyframes with up to three references to keep her face consistent, MiniMax H3 with image plus music in and motion plus lip-sync out, SeedVR2 7B to upscale drafts to 1344×768 without changing the motion, and HyperFrames for the old-movie look, titles and thumbnail.
Results and limits
The creator's trick: the intro is one 15-second take made in first/last-frame mode with a fixed end frame, so it lands exactly on the first frame of the next scene and there is no visible cut. The character now has her own character card so she can come back in later videos.
Worth knowing
This is the creator's own description in one post; no render times or settings are given. It shows a full music video pipeline with lip-sync is possible on a 12 GB consumer card.