🎤 Slip of a Tongue: Whisper Edition
Slip of a Tongue: Whisper Edition allows players to trigger server-side commands using their actual voice. Powered by the high-accuracy WhisperLib engine, your spoken words are processed locally and sent to the server for secure execution.
🔥 What's New in Whisper Edition?
-
High Accuracy: Powered by OpenAI’s Whisper, it understands natural speech, accents, and fast talkers far better than the old VoskLib.
-
Smart Noise Filtering: Automatically ignores background music, coughs, and non-speech sounds like [MUSIC] or *laughing*.
-
Zero Lag: Audio processing happens on a separate thread, so your FPS stays buttery smooth.
🛠 Installation
Client-Side (Required)
-
Install WhisperLib.
-
Install Slip of a Tongue: Whisper Edition.
-
Download a Model: Open the WhisperLib config in-game and download a model (we recommend base.en for the best performance-to-speed ratio).
Server-Side (Required)
-
Install Slip of a Tongue: Whisper Edition (WhisperLib is not required on the server).
-
Configure your word list in the server's config folder.
⚙️ Command Syntax
The server-side configuration uses a specific format to map voice triggers:
word:[command1, command2, %player%]Text to narrate
- word: The keyword or phrase to detect.
- [...]: List of commands (no
/ needed).
- %player%: Automatically replaced by the speaker's name.
- Text to narrate: The message sent to chat upon success.
Example: jump:[effect give %player% levitation 2 10]See you later!
🔒 Security & Privacy
- Local Processing: Speech-to-text happens entirely on your machine. No audio data ever leaves your computer.
- Server-Side Validation: The server only executes commands that are explicitly defined in its configuration, preventing players from spoofing unauthorized commands.
📜 Credits
- Engine: WhisperLib by MeroSsSany.
- Original Inspiration: VoskLib and Slip of a Tongue.
- Speech Model: Powered by OpenAI's Whisper.
[ Whisper Edition Branch ] [ Original GitHub Branch ]