Skip to content

Repository files navigation

Vyzrox Banner

⚡ Vyzrox Vision

Typing effect

Algorithmic intelligence, executed entirely in your browser.
Vyzrox utilizes Rust compiled to WebAssembly (WASM) to run a local neural core at near-native speeds.

Status Engine Privacy Latency

VisionArchitecturePrivacyQuick Start


👁️ Vision

In the era of cloud-dependent LLMs, user privacy and response latency are constantly compromised. Vyzrox flips the paradigm. By compiling Rust into WebAssembly and leveraging WebGPU, Vyzrox brings the neural processing core directly to the edge—your device.

No API calls, no hidden logging, no telemetry. What happens in your browser, stays in your browser.


⚙️ Architecture & Data Flow

Vyzrox operates on a strict isolated architecture. Below is the visual representation of how the engine processes data locally without ever touching an external server.

graph TD
    classDef user fill:#1a2030,stroke:#63b3ed,stroke-width:2px,color:#fff;
    classDef core fill:#0e1118,stroke:#b794f4,stroke-width:2px,color:#fff;
    classDef cloud fill:#2d3748,stroke:#fc8181,stroke-width:2px,color:#fc8181,stroke-dasharray: 5 5;
    classDef webgpu fill:#080b12,stroke:#48bb78,stroke-width:2px,color:#fff;

    U((👤 User Input)):::user -->|"Text / Image"| UI["React / Vite Interface"]:::user
    UI -->|ArrayBuffer| Hook["useVyzroxEngine hook"]:::user
    
    subgraph Browser Sandbox [Zero-Trust Browser Sandbox]
        Hook --> W["WebAssembly (WASM) Core"]:::core
        W <-->|"Hardware Acceleration"| GPU{"WebGPU Compute"}:::webgpu
        GPU -->|"Tokens / Inference"| LLM["Local Qwen/Vision Model"]:::core
    end
    
    LLM -.->|"Streaming Response <1ms"| Hook
    
    Hook -.-x|Blocked| C((Cloud Servers)):::cloud
    Hook -.-x|Blocked| T((Telemetry)):::cloud
    
    style Browser Sandbox fill:#080b12,stroke:#2d3748,stroke-width:1px
Loading

🛡️ Zero-Trust Privacy Model

Unlike traditional AI wrappers, Vyzrox guarantees total data sovereignty.

sequenceDiagram
    autonumber
    participant U as 👤 User
    participant V as ⚡ Vyzrox UI
    participant E as 🧠 Local WASM Engine
    participant C as ☁️ The Cloud

    U->>V: Enters prompt / Uploads image
    V->>E: Dispatches task to Web Worker
    Note over E: Executing Neural Inference<br/>via WebGPU & Rust
    E-->>V: Streams response (Zero Network Overhead)
    V-->>U: Displays Output
    
    rect rgb(255, 100, 100, 0.1)
    U-xC: No data transmitted
    V-xC: No API keys required
    E-xC: No telemetry logs sent
    end
Loading

✨ Features

⚡ WebAssembly Core

Rust compiled to Wasm executes in your browser at near-native speed. No servers, no round-trips.

🔒 Total Privacy

Your data never leaves your device. Zero network calls. Zero telemetry. You own every byte.

🚀 Zero Latency

Every response is analyzed in under a millisecond by the local neural core running at full CPU/GPU speed.


🚀 Quick Start (Initialize Core)

To run Vyzrox locally and experience zero-latency AI:

# 1. Clone the repository
git clone [https://github.com/adarshthakur9240/Vyzrox.git](https://github.com/adarshthakur9240/Vyzrox.git)

# 2. Navigate to the UI directory
cd vyzrox-ui

# 3. Install dependencies
npm install

# 4. Initialize the local dev server
npm run dev

Note: Ensure your browser supports WebGPU and WebAssembly for optimal inference speeds.


🛠 Tech Stack

  • Frontend: React (Vite), TailwindCSS, Framer Motion
  • Engine: WebAssembly (WASM), WebGPU, Web Workers
  • Routing: React Router DOM
  • Styling: Custom CSS Variables (Cyberpunk / Dark Glassmorphism)

Built by Adarsh's Team

© 2026 Vyzrox Vision Suite. All operations are local.

About

Algorithmic intelligence, executed entirely in your browser. Vyzrox utilizes Rust compiled to WebAssembly (WASM) to run a local neural core at near-native speeds.

Topics

Resources

Stars

Watchers

Forks

Releases

Packages

Contributors

Languages