voice-replacementlisted

Use when the user asks to CHANGE, REPLACE, RE-VOICE, or DUB the voice on one or more EXISTING videos/scenes in a piece — a deliberate, user-triggered step AFTER the video exists ("change the voice", "give it a different voiceover", "redo the voice", "dub this", "clone my voice over it", "new narrator"). This is NOT initial generation (that keeps native audio via voiceover-production). It transcribes the target scenes, asks whether to clone the existing voice or pick a new one, lip-syncs the sections where a character speaks on camera, and mutes + re-voices the rest. A standalone entry point with its own trigger.
Nagellabs/libi · ★ 2 · AI & Automation · score 72

Install: claude install-skill Nagellabs/libi

# Voice Replacement — re-voice an existing video The user has a video (one or more scenes) and wants a **different voice** on it — cloned from the original speaker or a brand-new voice. This is a separate flow from generation: by default a piece keeps its native / `@Audio1`-carried audio (`voiceover-production` owns that). Run THIS skill only when the user explicitly asks to change/replace the voice on footage that already exists. **Two hard principles** (carried over from `voiceover-production`): - **MUTE, never delete** the original audio — `libi.audio_update_clip({ clipId, enabled:false })`, not `audio_remove_clip`. The original stays on the timeline and is one-click-toggleable (per-clip speaker icon); `audio_relink_overlay` is the only recovery if it was deleted. - **Cover the actual speech.** Each new voice segment must cover what was actually said in that scene (sized from the transcript) — not a short summary that under-fills and leaves the speaker silent. `≤` scene duration is a guardrail against overflow, not the target. ## Workflow ### 0. Scope — which scenes? Ask (or confirm) which scenes get the new voice: **all** of the piece's video overlays, or a **subset** (e.g. "just the hook and verdict"). List the scenes you'll touch back to the user before spending anything. ### 1. Transcribe each target scene (reuse if it exists) For each target scene's video file, get a transcript **with word-level timing** — this is the coverage anchor (it tells you WHO spe