Vocalizer

Overview

Vocalizer is for making typed words sing your melody without pretending to be a real human vocalist. Type the line, capture or play a MIDI melody, hit Generate, and the bundled neural TTS voice follows the notes with formant-preserving pitch correction. You can audition the result in the plugin, then drag the rendered WAV straight onto an audio track. It is great for hooks, vocal chops, bass music textures, weird robot lines, and talkbox/autotune-style ideas. It runs offline and ships five voices (CORI, LJ, KRISTIN, NORMAN and JOHN), all trained on public-domain recordings, and it is free/open source under GPLv3. Right now it is macOS Apple Silicon only.

Features

• Type text, capture MIDI, Generate
• Offline neural TTS using Piper / ONNX
• Formant-preserving pitch correction with TD-PSOLA
• Five bundled voices, all trained on public-domain recordings
• Add your own Piper voices in ~/Documents/Vocalizer/Voices/
• Drag rendered WAVs straight into your DAW
• Native Apple Silicon with saved plugin state
• Free and open source under GPLv3

New here? The install guide walks through AU / VST3 setup on Mac and Windows.