menogt

Voiceframe — AI skill for Claude Code

AI community

Claude skill that turns a reference motion-graphics video and your voiceover into timed AI video prompts (Omni Flash, Veo, Kling, Seedance, Sora, Runway), plus a correctly spelled text-overlay video s.

How to install Voiceframe

This entry records only its repository, not the path inside it, so there is no exact command to give. Open menogt/voiceframe and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Voiceframe does

Claude skill that turns a reference motion-graphics video and your voiceover into timed AI video prompts (Omni Flash, Veo, Kling, Seedance, Sora, Runway), plus a correctly spelled text-overlay video synced to your voice.

Alternatives in AI

  • Gsap Skills — Official AI skills for GSAP 1.7k ★
  • Make Prompt Seedance2 — 一句话:试错成本越来越高,提示词的重要性从来没有像今天这样大 418 ★
  • Open AI Design Agent — Open-source AI design agent — alternative to Lovart AI, Runway Agent, Luma Labs Agent, Krea Agent, Pika Agent 356 ★

README

Voiceframe by Meno

**v1.0** · by Menaka Perera · MIT License

Give Claude a motion-graphics video you like and your own voiceover. You get AI video prompts that copy the look, with every beat timed to your voice, plus a perfectly spelled text layer that drops straight onto your timeline.

**Before:** a reel you like and an MP3 of your narration. **After:** a style bible, a beat sheet timed to the word, one copy-paste prompt per clip for Omni Flash / Veo / Kling / Seedance / Sora / Runway, and a text-overlay video (transparent MOV plus a green-screen MP4) already synced to your voice.

Install

  1. Download voiceframe-by-meno.zip.
  2. In Claude, open Settings, go to the Skills section (under Capabilities) and upload the zip.
  3. Turn on code execution (the skill runs ffmpeg and a small renderer).

What to attach

  • Reference video (the style). Under 60 seconds works best. Reels, explainers and ads all work.
  • Your voiceover audio (best: timing comes from your actual voice) and/or your script. A captions file (.srt or .vtt) from CapCut, Premiere or YouTube also works as the script.

Then say something like *"recreate this style for my voiceover"*. Claude asks one question with defaults (ratio, test or full, text layer, target model). Reply "go" to accept them.

Scripts can use tags. They are optional:

[VO] It's three in the morning somewhere, and someone is awake.
[VIS] AI — a phone screen is the only light in a dark room
[TEXT] 3 a.m.
[BEAT]
[VIS] STOCK — empty street at night

`[VO]` spoken · `[VIS]` / `[VIS] AI` / `VEO` / `DATA` → Claude writes a prompt · `[VIS] STOCK` / `CANVA` / `Title card` → goes to your sourcing list · `[TEXT]` → exact words for the text layer · `[BEAT]` → a pause and a good place to cut.

What you get

  1. Style bible: palette with hex codes, typography, motion, transitions and pacing, taken from the reference (style only).
  2. Beat sheet: every beat with its time on your voiceover and the emph