Communities

R.E.P.O.Speech2TTS1.2.0

You are viewing a potentially older version of this package. View Latest Version

Install

Date Uploaded

8 months ago

Downloads

2.7K

Size

55MB

Dependency string

NocturnalLyfe-Speech2TTS-1.2.0

Download

Details Required (1)Versions

Speech-2-TTS Mod by NocturnalLyfe

Transform your voice chat into hilarious text-to-speech! Uses completely offline speech recognition.

Features

Real-time voice -> text -> TTS conversion
Your actual voice is muted when STT is enabled (only TTS is broadcast)
Toggle on/off with a single keypress
Text appears in game chat for all players to see
Flexible model support - use small, medium, or large models
No performance impact

Installation

Automatic (Mod Manager)

Install via Thunderstore Mod Manager or r2modman
Launch the game

Manual

Install BepInEx if not already installed
Extract the mod folder to REPO/BepInEx/plugins/
Launch the game

Usage

Controls

F7 - Toggle Speech-to-Text mode ON/OFF

Modes

STT Enabled (Default):

Your voice is converted to text
Text is broadcast via in-game chat as TTS
Your actual voice is muted (others only hear TTS)
Perfect for robot voice fun!

STT Disabled:

Normal voice chat
Your real voice is transmitted
No text-to-speech conversion

How It Works

Speak into your microphone
Vosk speech recognition transcribes your voice
Text is sent to the game's chat system
Game's TTS speaks your message to all players
Everyone sees and hears your transcribed speech!

Configuration

No configuration needed! Just press F7 to toggle.

Using Different Models

The mod automatically detects which model you have installed (in priority order):

vosk-model-small-en-us-0.15 (~40MB) - Default, fastest, good accuracy
vosk-model-en-us-0.22 (~1.8GB) - Better accuracy, slower
vosk-model-en-us-0.42-gigaspeech (~2.3GB) - Best accuracy, slowest

To Use a Different Model:

Download your preferred model from https://alphacephei.com/vosk/models
Extract to BepInEx/plugins/NocturnalLyfe-Speech2TTS/model/
Keep the original folder name (don't rename it)
The mod will automatically detect and use it!

Tip: You can have multiple models installed - the mod will use the first one it finds in the priority order above.

Warning: Changing to larger model will cause longer launch of REPO and requires a stronger device to run.

Troubleshooting

Mod not working:

Check BepInEx/LogOutput.log for errors
Ensure all DLLs are in the plugin folder (libvosk.dll, libgcc_s_seh-1.dll, libstdc++-6.dll, libwinpthread-1.dll)
Verify at least one model folder exists in the model/ directory

Low accuracy:

Speak clearly and at a moderate pace
Check microphone quality
Consider upgrading to vosk-model-en-us-0.22 or vosk-model-en-us-0.42-gigaspeech for better accuracy

No audio transmission:

This is normal when STT is enabled! Your voice is muted and replaced with TTS
Press F7 to disable STT for normal voice chat

Wrong model being used:

Check the log file to see which model was loaded
Remove unwanted models from the model/ folder
Models are loaded in priority order (small -> normal -> gigaspeech)

Technical Details

Default Model: vosk-model-small-en-us-0.15 (~40MB)
Speech Recognition: Offline Vosk engine
Sampling Rate: 48kHz
Language: English (US)
Supported Models: Any Vosk English model compatible with the API

Credits

Mod by NocturnalLyfe
Vosk Speech Recognition: https://alphacephei.com/vosk/
BepInEx: https://github.com/BepInEx/BepInEx

Support

Report issues on GitHub or the Thunderstore page. https://github.com/NocturnalLyfe/REPO-Speech2TTS

Changelog

See CHANGELOG.md for version history.