Generate mumble / gibberish dialogue sounds for game characters.
Not TTS. It turns dialogue rhythm into cute character blips, WAV files, and subtitle reveal timing.
▶ Live Demo · ♪ 12-clip Showcase · v1.5.0 Release · ⤓ Godot addon zip · § Engine docs
中文 · English
Composer on the left for typing the line and tweaking emotion / style / intensity. Live preview on the right with the blip timeline and analysis grid. Nine character presets sit below, with the Sound Lab advanced drawer underneath.
Mumble Voice Lab is a browser-based character mumble / gibberish voice generator for games. Type a line, choose a character, tune emotion and speaking style, preview instantly, then export deterministic WAV + mumble-voice-lab/schedule JSON.
It is built for cozy RPGs, indie games, visual novels, creature games, NPC dialogue prototypes, and any project that needs expressive character dialogue sounds without real speech synthesis.
Tip
Placeholder lines can mix languages, e.g. 你好 adventurer, ready for today's quest? — the engine times blips from character density and punctuation.
|
Not TTS |
WAV + schedule JSON |
|
Deterministic output |
Preset + emotion + style |
|
Reveal events |
Game-ready assets |
| Unity alpha | Godot Windows-first | Generator dock |
|---|---|---|
 Local UPM package. It still depends on local Node/npm and calls npx tsx scripts/mvl.ts to generate assets. |
 Godot addon 0.2.0. On Windows it defaults to the bundled mvl-renderer-win-x64.exe, so normal users do not need Node. |
 Type dialogue, choose preset / emotion / style, then generate WAV + .mumble.json + MumbleDialogueClip .tres. |
Note
The runtime boundary is intentional: engines play generated assets. MumbleVoicePlayer syncs subtitles and typewriter UI from revealEvents. Player-entered free-text synthesis at runtime is not part of this release.
 1. Enter dialogue Write one NPC line in Chinese, English, or mixed text. |
 2. Analyze rhythm Estimate pseudo-syllable events from text length, punctuation, phrases, and language features. |
 3. Generate mumble voice Combine preset and expression settings into character-like blips. |
 4. Export assets Write WAV and schedule JSON; batch renders also produce a manifest. |
 5. Sync subtitles revealEvents provide exact timing for UI text reveal.
|
 6. Play in game Unity / Godot runtime plays audio and dispatches reveal events. |
| Scenario | Use it |
|---|---|
| Web tool | Open the Live Demo, type a line, preview, and export. |
| CLI | npm run mvl -- render --text "Good morning, traveler! Ready?" --preset cute-npc --out-dir out |
| Batch | npm run mvl -- batch --input dialogue.csv --out-dir out |
| Unity alpha | Add integrations/unity/com.nightt5879.mumble-voice-lab as a local UPM package, run npm install, then open Tools > Mumble Voice Lab. |
| Godot 0.2.0 | Download the Godot zip from the release, or copy integrations/godot/addons/mumble_voice_lab into a Godot 4.6 project and enable the plugin. |
Sound Lab · advanced parameters (click to expand)
The in-app Sound Lab drawer exposes 6 parameter groups (pitch · rhythm · roughness · brightness · continuity · level) plus global toggles like pitchFallAtEnd. Every override is preserved in the exported schedule JSON, so the same line reloads to identical output.
Full parameter table lives in docs/integrations.md.
 |
 |
 |
|---|---|---|
| Platform coverage | Feedback channel | Release positioning |
| Godot Windows verifies bundled renderer, headless tests, and manual playback. macOS / Linux do not include a bundled renderer yet; use the Node CLI fallback for development. | Complex Unity / Godot projects may expose path, import, export, or runtime issues. Please open issues with repro steps. | Web tool and export protocol stable · Unity alpha · Godot windows-first store-ready candidate. Final Asset Library acceptance depends on official review. |
| Version | Focus |
|---|---|
V1.1.0 |
Smoother blip transitions, envelopes, fades, and connected playback. |
V1.2.0 |
Cozy sticker visual system, character avatars, and crowd chatter. |
V1.3.0 |
"My Presets" custom voice saving, export, and import. |
V1.4.0 |
CLI renderer, schedule JSON 1.0, Unity local UPM alpha, and Godot preview. |
V1.5.0 ← current |
Godot 0.2.0 Windows-first: bundled renderer, .tres dialogue resources, headless tests, and Asset Library materials. |
Release notes live in CHANGELOG.md. Engine setup and QA steps live in docs/integrations.md.
# install deps and start Vite
npm install
npm run dev# production build (outputs to dist/)
npm run build# regenerate the 12 listening samples for the showcase page
npm run samplesCopyright 2026 nightt5879.
Code is released under the Apache License 2.0.
Audio files, JSON schedules, and other outputs generated with Mumble Voice Lab may be used freely in personal, commercial, and open-source game projects.
cozy handmade tool · 2px ink · sticker shadow · 1 page