voice-act

Render character voiceover MP3s from scene prose via ElevenLabs MCP.

Updated May 9, 2026

Search Tags:#storytelling #transcripts #elevenlabs #tts #voiceover #audio rendering #character casting

Ready to run

npx skills add https://github.com/PetitChu/holocron --skill voice-act

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: voice-act
Source: https://github.com/PetitChu/holocron/tree/main/skills/voice-act
Command: npx skills add https://github.com/PetitChu/holocron --skill voice-act

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill turns finished scenes and lines into character voiceover audio so you can audition, confirm, and re-render voices without rewriting your manuscript.

Core Features & Use Cases

Voice audition & casting: Render the same sample across candidate voices for a character so you can choose the best match.
Soundbite voice confirmation: Produce a longer “ID photo” render for a character after you lock the voice assignment.
Scene and line audio rendering: Generate multi-speaker scene MP3s or single-line re-takes, while writing transcript sidecars and a render log.

Quick Start

Run /holocron:voice-act scene 3 2 to render the finished scene into stitched MP3 audio with transcript sidecars.

Frequently Asked Questions about voice-act

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I render character voiceover audio from finished scene prose?▼

To render character voiceover audio, this Skill processes finished scene prose using the ElevenLabs MCP for TTS, generating stitched MP3 audio outputs alongside transcript sidecar files for multi-speaker scenes.

Can I audition multiple voices for a single character before casting?▼

Yes, you can audition character voices by rendering the same text sample across multiple candidate TTS voices, allowing you to compare outputs and choose the best match before locking the cast assignment.

Do I need an ElevenLabs MCP server to generate scene MP3s and transcripts?▼

Yes, access to the ElevenLabs MCP server is required. The Skill uses this connection to produce TTS audio, writing MP3 files under inserts/audio with accompanying transcript sidecar files.

What is the best way to produce single-line re-takes for multi-speaker scenes?▼

The best way to produce single-line re-takes is to run a targeted render command on the finished prose. The Skill generates the specific line audio and updates the transcript sidecar without rewriting your manuscript.

How does character casting context work when generating voiceover audio?▼

Character casting context works by reading project files like world/voice-cast.md alongside character and narration context, ensuring the TTS audio rendering applies the correct assigned voices to each speaker.