Kokoro TTS
Script to Voice Generator - Kokoro TTS is a downloadable Windows desktop tool by Reactorcore that converts formatted .txt or .md script files into voiced audio using Kokoro ONNX, a local text-to-speech engine. It runs entirely offline, requiring no API keys, cloud account, usage limits, or per-use costs after setup. Each speaker in a script can be assigned a voice, pitch, speed, volume, and audio effects. The tool includes 49 built-in voices across 8 languages, including two English variants, Mandarin Chinese, Spanish, French, Hindi, Italian, and Brazilian Portuguese. A Voice Blender allows interpolation between two or three base voices at custom ratios, with saved blends appearing alongside built-in voices. It also offers 13 audio effects with Off, Mild, Medium, and Strong presets, plus FMSU corruption and Reverse toggles. The generator produces per-line clean and effects-processed clips, merged audio files with smart pacing, loudness-normalized merged output for podcasts, raw merged output for media and games, and a reference text file listing filenames, line numbers, and spoken content. It supports sound effect events in the merge timeline, character profiles saved between sessions, and punctuation-based pause timing. System requirements note Windows 11 as tested; Windows 10 is untested and no Linux or macOS build is available. It uses DirectML to automatically run on GPU when available or CPU otherwise.
- Category
- Video
- Pricing
- Unknown
Problem solved
Creating voiced audio from scripts often requires cloud TTS APIs, API keys, usage limits, or manual editing. This tool solves that by converting formatted scripts into per-character voiced audio locally on Windows, with voice assignment, audio effects, and merged output generation, without cloud costs or limits.
Use cases
- Generate game dialogue for multiple characters from a script
- Produce audio dramas with per-character voices and effects
- Create audiobook or narration tracks from .txt or .md manuscripts
- Generate ambient narration or meditation audio
- Build voice banks with blended character voices
- Create character one-liners and single-word exclamations using Yell Impact mode
- Produce podcast-style loudness-normalized merged audio
- Prototype voice acting for animations, videos, or interactive media
Audience
- Content Creators
- Developers
- Hobbyists
Pros
- Fully offline operation with no API keys, cloud accounts, or usage limits
- 49 built-in voices across 8 languages plus extensive voice blending
- Per-character voice, pitch, speed, volume, and 13 audio effects
- Outputs both per-line clips and merged audio, including loudness-normalized and raw versions
- Uses DirectML to automatically run on GPU when available or CPU otherwise
Cons
- Windows-only; no Linux or macOS build
- Windows 11 is tested while Windows 10 is untested
- Requires local installation and setup; not a hosted or browser-based service
- Pricing is 'name your own price' on itch.io, and licensing details are not stated on the homepage
- Voice and effect quality depends on the local Kokoro ONNX model and system resources
Official website: https://reactorcore.itch.io/script-to-voice-generator-kokoro-tts
Canonical LambdaBase page: https://www.lambdabase.com/tools/kokoro-tts