Z-AI:Personal Assistant

Z-AI is a high-performance, multi-modal personal assistant built in Rust. Unlike traditional assistants, Z is designed with a "Toxic, Savage, and Unfiltered" personality, offering a unique (and often roasting) interaction experience.

It supports voice and text input, integrates with multiple LLM providers, and can control system media across macOS, Linux, and Windows.

Features

• Multi-Modal Input: Seamlessly switch between voice (Wake Word activated) and keyboard input.
• Aggressive Personality: Powered by custom system prompts to be sarky, toxic, and brutally honest.
• Triple-Provider LLM Support:
  • Google Gemini (Primary: 2.0 Flash)
  • OpenAI (gpt-4o-mini)
  • Groq/xAI (Llama 3.3 / Grok-Beta)
• High-Fidelity Voice:
  • STT: Fast transcription via Groq (Whisper-large-v3).
  • TTS: Premium voice synthesis via ElevenLabs, with local system fallbacks.
• Universal Music Control: Control Spotify or YouTube Music via voice commands (PLAY, PAUSE, STOP).
• Cross-Platform: Tailored commands for macOS (osascript), Linux (playerctl), and Windows (powershell).

Getting Started

Prerequisites

• Rust: Install Rust (Edition 2021).
• System Dependencies:
  • Linux: alsa development files (e.g., libasound2-dev), playerctl.
  • macOS: Native support.
  • Windows: PowerShell.

Environment Setup

Create a .env file in the root directory and add your API keys:

GEMINI_API_KEY=your_gemini_key
OPENAI_API_KEY=your_openai_key
GROQ_API_KEY=your_groq_key
ELEVEN_LABS_API_KEY=your_elevenlabs_key

Installation

1. Clone the repository:

   git clone https://github.com/HANZSEIO/Z-ai_Rust.git
   cd Z-ai_Rust
   

2. Run the application:

   cargo run --release

Usage

1. Wake Word: Say "Z", "Zee", or "Zed" to activate the assistant.
2. Active Mode: Once activated (indicated by a sound and UI prompt), you have 60 seconds to speak.
3. Music Control: Try saying:
   • "Z, play some Phonk"
   • "Stop the music"
4. Text Mode: Simply type your message in the terminal at any time.
5. Exit: Type exit or quit.

Architecture

• src/main.rs: Orchestrates the async event loop using tokio, managing concurrent audio listening and terminal input.
• src/core/audio.rs: Handles low-level audio stream management (cpal), WAV encoding (hound), and STT/TTS API calls.
• src/core/cloud_api.rs: Manages the LLM logic, system prompts, and fallback mechanisms between providers.

Disclaimer

Z-AI is configured to be toxic and unfiltered. It is intended for entertainment purposes and "toxic friend" simulations. But you can configure manualy Use at your own risk :).

Version

Z-ai pacth will be updated on my wiki, maybe at 10 june. so wait for new patch :).

Built with attitude.

by HANZS -