VoiceStudio

Overview

VoiceStudio is a local-first voice cloning + text-to-speech studio powered by Qwen. I built it because I wanted a workflow closer to “studio mode” (think ElevenLabs Studio), but running on my own machine with more control.

What it solves

Long scripts get automatically split into paragraphs for more stable generation.
You can assign different speakers per paragraph (podcast/conversation style).
Everything runs as a queue so you can iterate, review, and export cleanly.

Features

Paragraph queue workflow
Per-paragraph speaker assignment
Timeline playback and final audio export
Voice preset library (create / rename / delete)
Backend health checks + local runtime auto-start orchestration

Demo

Screenshots

Stack