flirtatiousAI 1.0

A playful AI companion built with modern web technologies. Features real-time streaming responses, intelligent rate limiting, and a polished cyberpunk-inspired UI.

Features

• SvelteKit 2 with Svelte 5 runes for reactive state management
• Streaming responses for real-time conversation flow
• Smart rate limiting (3 requests per 5 minutes per IP)
• Toast notifications replacing intrusive alerts
• Hugging Face Inference with Llama 3.2 fallback
• TypeScript with strict type checking
• Tailwind CSS 4 for responsive design
• Vercel-ready deployment

Tech Stack

• Svelte 5 with runes ($state, $derived)
• SvelteKit 2
• TypeScript
• Tailwind CSS 4
• Vercel adapter

Quick Start

# Clone
git clone https://github.com/MaDrCloudDev/flirtatiousAI.git
cd flirtatiousAI

# Install
pnpm install

# Configure
echo "HF_ACCESS_TOKEN=your_token_here" > .env

# Run
pnpm dev

Get your token from Hugging Face Settings (requires inference.serverless.write permissions).

Project Structure

src/
├── lib/
│   ├── components/
│   │   └── Toast.svelte      # Toast notification system
│   ├── notifications.svelte.ts # State management for toasts
│   └── types.ts              # Shared TypeScript types
├── routes/
│   ├── api/flirt/
│   │   └── +server.ts        # Streaming API endpoint
│   ├── +layout.svelte        # Root layout
│   └── +page.svelte          # Main chat interface
├── app.css                   # Global styles
└── app.html                  # HTML template

Scripts

API

POST /api/flirt

Request:

{ "context": "Hi, I'm Alex..." }

Response: Server-Sent Events stream with JSON chunks

Rate limited to 3 requests per 5 minutes per IP.

License

MIT