Script to Voice Generator - Kokoro TTS

Script to Voice Generator - Kokoro TTS

Turn formatted scripts into fully voiced audio — local, offline, no API keys, no limits.

## Description

Script to Voice Generator converts formatted `.txt` or `.md` script files into fully voiced audio using **Kokoro ONNX** — a high-quality, local text-to-speech engine that runs entirely on your machine. No API keys. No cloud account. No usage limits. No costs. Everything is offline.

Each speaker in your script gets their own voice, pitch, speed, and audio effects. Assign Radio, Reverb, Distortion, Telephone, Robot Voice, Cheap Mic, Underwater, Megaphone, Worn Tape, Intercom, Alien Voice, Cave, or Pitch Shift effects per character at Off / Mild / Medium / Strong levels. Two bonus toggles — FMSU (brutal digital corruption) and Reverse — round out the toolkit. Combine effects freely for distinct character identities.

Kokoro includes **49 built-in voices across 8 languages** — English (US and UK), Mandarin Chinese, Spanish, French, Hindi, Italian, and Brazilian Portuguese — and that's just the starting point. The built-in **Voice Blender** lets you interpolate between two or three base voices at custom ratios to create entirely new characters. With 49 voices, there are over **19,000 unique blend combinations** before you even touch the ratio sliders — and since every blend is a continuous spectrum (not a fixed preset), the actual space of distinct-sounding voices you can craft is effectively limitless. Blended voices are saved and appear in every speaker panel, ready to use like any built-in voice.

**System requirements:** Windows 11 (tested). Windows 10 untested, try at your own risk. No Linux or macOS build available. Uses DirectML — automatically runs on your GPU if available, CPU otherwise, no setup required.

**The generator produces:**
- Individual audio per each spoken line — clean (TTS only) and effects-processed versions
- All (effects-enabled) clips merged into a fully edited and smartly paced audio file, not normalized (true audio, better for media/games)
- All (effects-enabled) clips merged into a fully edited and smartly paced audio file, loudness normalized (even audio, better for podcasts)
- Reference .txt file with filenames, line numbers, and spoken content for every clip

**Features:**
- 49 built-in Kokoro voices across 8 languages (English US/UK, Chinese, Spanish, French, Hindi, Italian, Portuguese)
- Voice Blender — 19,000+ blend combinations from 49 base voices, with continuous ratio control; saved blends work like any built-in voice
- Multilingual TTS — correct language phonemization selected automatically per voice
- Per-character voice, pitch, speed, volume, and audio effects
- 13 audio effects — most with Off / Mild / Medium / Strong presets
- 2 bonus toggles: FMSU (brutal corruption) and Reverse
- Yell Impact mode for punchy single-word exclamations
- Inner thoughts filter (Whisper, Dreamlike, Dissociated presets)
- Sound effect events (play/stop/loop) placed in the merge timeline
- Smart merged audio with configurable punctuation-based pause timing
- Loudness-normalized and raw merge outputs
- Per-line clip files (clean and effects) for editing flexibility
- Character profiles saved automatically between sessions
- Parse log with line-by-line error reporting
- Included example scripts and AI prompt templates to get started fast
- Fully offline — no internet connection required after setup

**Ideal for:**
- Game dialogue
- Character oneliners
- Audio dramas
- Audio books
- Ambient narration
- Meditation audios
- Voice banks

...and any project where you want to convert a written script into audio.
Fully automated, high quality, and no limits.

*Includes in-depth guides covering script writing for TTS and the full audio effects pipeline.*

---

Check out everything else I do: ✨🚀
https://linktr.ee/reactorcore