Levi / frontend /src /components /AboutPage.tsx
Manish Kumar
chnage model
0a7f861
Raw
History Blame Contribute Delete
4.28 kB
import type React from 'react';
import { Info, Cpu, Code, BookOpen, HardDrive, Share2 } from 'lucide-react';
export const AboutPage: React.FC = () => {
return (
<div className="flex-1 overflow-y-auto bg-background p-6 lg:p-8 space-y-8 select-none">
<div className="flex items-center gap-2 border-b border-border pb-4">
<span className="p-2 bg-primary/10 rounded-xl text-primary">
<Info size={20} />
</span>
<div>
<h1 className="text-2xl font-bold tracking-tight">About Antigravity AI Coder</h1>
<p className="text-xs text-muted-foreground">Learn about the architecture and technologies running under the hood.</p>
</div>
</div>
<div className="max-w-3xl space-y-6">
<div className="bg-card border border-border p-6 rounded-2xl space-y-4">
<h2 className="text-base font-bold flex items-center gap-2">
<Cpu size={18} className="text-primary" />
Hybrid Local/Cloud Architecture
</h2>
<p className="text-sm text-muted-foreground leading-relaxed">
Antigravity Coder is built on a cascading inference pipeline. It is configured to run the lightweight <strong>SmolLM2-360M-Instruct</strong> model locally in GGUF format using <code>llama-cpp-python</code>.
</p>
<p className="text-sm text-muted-foreground leading-relaxed">
If the application is deployed on cloud hosts with restricted memory boundaries (e.g. Render Free, Railway Starter), the backend automatically and transparently switches to the <strong>Hugging Face Serverless Inference API</strong>, guaranteeing high-performance execution without needing a local GPU or extensive memory budgets.
</p>
</div>
<div className="grid grid-cols-1 md:grid-cols-2 gap-4">
<div className="bg-card border border-border p-5 rounded-2xl space-y-2">
<h3 className="text-sm font-bold flex items-center gap-1.5">
<Code size={16} className="text-primary" />
SmolLM2 Engine
</h3>
<p className="text-xs text-muted-foreground leading-relaxed">
Leverages the SmolLM2 360M Instruct model, optimized for fast inference on CPU with excellent instruction-following and code assistance capabilities.
</p>
</div>
<div className="bg-card border border-border p-5 rounded-2xl space-y-2">
<h3 className="text-sm font-bold flex items-center gap-1.5">
<HardDrive size={16} className="text-primary" />
GGUF Quantization
</h3>
<p className="text-xs text-muted-foreground leading-relaxed">
Quantized in Q4_K_M (4-bit quantization). Reduces the weights size to under 400MB and fits comfortably within 1.5GB system RAM, providing blazing-fast local processing speed.
</p>
</div>
<div className="bg-card border border-border p-5 rounded-2xl space-y-2">
<h3 className="text-sm font-bold flex items-center gap-1.5">
<BookOpen size={16} className="text-primary" />
Monaco Code Editor
</h3>
<p className="text-xs text-muted-foreground leading-relaxed">
Integrates the VS Code core editor (Monaco), delivering built-in code formatting, syntax highlighting, search/replace, and advanced editing features directly inside the Playground.
</p>
</div>
<div className="bg-card border border-border p-5 rounded-2xl space-y-2">
<h3 className="text-sm font-bold flex items-center gap-1.5">
<Share2 size={16} className="text-primary" />
Cloud Container Ready
</h3>
<p className="text-xs text-muted-foreground leading-relaxed">
Dockerized with independent backend and frontend packages, fully prepared for automated pipeline deployment on Railway, Render, and Fly.io.
</p>
</div>
</div>
<div className="text-center text-xs text-muted-foreground pt-4">
Antigravity AI Coding Assistant © 2026. Made with Google DeepMind Advanced Agentic Coding.
</div>
</div>
</div>
);
};