Can I use voices other than the default Irene voice?

Yes, the skill supports a 'Custom Voice' mode using built-in speakers like Sohee, and you can provide your own reference WAV files for custom voice cloning.

How long does the audio generation take?

On Apple Silicon with MPS acceleration, it typically takes 15-30 seconds per 300-character chunk. A full technical report usually takes between 10 and 30 minutes to synthesize completely.

Does this skill require a paid API like ElevenLabs?

No, this skill runs the Qwen3-TTS model locally on your machine, which means there are no per-character costs or cloud API subscriptions required.

What hardware is required for this TTS skill?

The skill is optimized for Apple Silicon (M1/M2/M3) using MPS acceleration, though it can run on any system capable of supporting the Qwen3-TTS Python environment.

Korean Markdown-to-Speech (Irene TTS)

Name: Korean Markdown-to-Speech (Irene TTS)
Author: jaminitachi

byjaminitachi

0•

コンテンツ管理

Converts Markdown documents into natural Korean speech audio using high-fidelity voice cloning and the Qwen3-TTS pipeline.

This skill automates the transformation of lengthy Markdown reports and documentation into high-quality Korean audio narrations. By leveraging the local Qwen3-TTS model with Apple Silicon (MPS) acceleration, it allows users to generate professional-sounding speech featuring the 'Irene' voice clone or custom speakers without incurring cloud API costs. It is an ideal solution for developers using SuperClaw who wish to consume generated reports while multitasking or commuting, featuring intelligent document chunking and dry-run validation to ensure perfect synthesis every time.

主な機能

01High-fidelity Korean voice cloning using the Irene reference profile

02Local execution on Apple Silicon (MPS) for cost-free, private synthesis

03Dry-run verification to preview character counts and section structures

04Dual-mode support for Voice Cloning (X-vector) and Custom Speaker profiles

05Automatic Markdown parsing with intelligent section-based chunking

060 GitHub stars

ユースケース

01Converting long SuperClaw-generated reports into audio for hands-free consumption

02Creating warm-toned Korean voiceovers for internal presentations and reports

03Generating narrated versions of technical documentation or project READMEs

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add jaminitachi/superclaw tts

For use in Claude.ai and ChatGPT

Download Skill