music-video-subtitle-generator
For musicians, video creators, and social-media editors producing AI music videos or emotional short films with lyric typography. Users provide music, lyrics, references, characters, typography direction, mood, or target platform. The Skill analyzes beat and vocal timing, separates character, scene, and text references, designs beat-reactive spatial typography, decomposes long works into connected shots, audits prompts, and routes generation for H3 or other video tools. It outputs MV concepts, s
Security Assessment
About music-video-subtitle-generator
Music Video Subtitle Generator (Music Aesthetics MV) creates, revises, audits, and routes generation for AI music videos and emotional short films where music, lyrics, on-screen typography, references, rhythm, performance, and camera language are designed together. It targets the problem that lyric-driven MVs need beat-synced typography and coherent multi-shot structure that ordinary caption or editing tools do not provide, treating on-screen text as a designed visual layer rather than plain subtitles.
The workflow starts with a pre-flight lock (aspect ratio and resolution options from vertical Shorts to cinematic 21:9, target duration, and music-window selection) and a lyric-lock step so the on-screen text matches the performed words verbatim. It then builds a creative contract (genre, BPM feel, vocal mode, mood), assigns each reference image a single narrow role (character card, scene card, typography card) to prevent cross-contamination, and applies preset visual grammars (for example Dark-pop, Cyber-grunge, Trap) with hard-cut editing and beat-reactive text. Its defining capability is a multi-shot stitching protocol for videos longer than the model's single-generation limit: it decomposes a long piece into 2-5 second beat-mapped shots, uses tail-frame/head-frame continuation and match cuts, aligns everything to one global master audio track, and enforces five continuity locks (vocal/lip, rhythm, aesthetic/color, transition/motion, typography). A checklist and canvas-delivery step finish the process.
It is aimed at musicians, video creators, and social-media editors producing stylized MVs and subtitle-driven music visuals. It is explicitly not for ordinary subtitle burn-in, generic editing, non-music ads, or simple single-clip requests, and it requires the MiniMax Hub agent's canvas and generation/routing tools rather than third-party shell scripts or external APIs.
FAQ
What makes it different from adding subtitles to a video?
It treats on-screen text as a designed, beat-reactive spatial typography layer that matches the performed lyrics word-for-word, not as ordinary caption burn-in; it also plans shots, references, and camera language together with the music.
How does it handle videos longer than the model's limit?
For clips over the single-generation limit (about 15 seconds) it uses a multi-shot stitching workflow: break the piece into 2-5 second beat-mapped shots, lock one global master audio track, use tail/head-frame continuation and match cuts, and align cuts to the beat grid.
How are reference images used?
Each reference is given one narrow role - character card, scene card, or typography card - and must not contaminate the others, so text references never contain characters or scenes and vice versa.
What are the continuity safeguards?
Five stitching locks: vocal and lip continuity, rhythm and beat matching, aesthetic and color grading, transition and motion continuity, and typography motion stitching, checked before delivery.
What environment does it need?
The MiniMax Hub agent with its canvas text nodes and image/video/music/editing agents; it does not assume third-party tools, shell scripts, browser plugins, or external generation APIs.
Install music-video-subtitle-generator
Quick Setup:
- Copy the skill folder to
.claude/skills/ - Claude will automatically detect and use the skill
Repository
minimax-ai/minimax-h3