AI图片生成 - AI绘画生成器 作者： Mindix

Generate AI Images turns text prompts into pictures entirely on your own device.

Inference runs in your browser on WebGPU — there are no servers, no sign-in, and nothing is ever uploaded. Your prompts and the images you create stay private, on your machine.

How it works
The model (SDXL Lightning, 1024×1024) downloads once into the browser's local storage and then runs fully offline. After the first download you can generate without any network connection.

Features
• Text-to-image with SDXL Lightning on WebGPU
• 100% local and private — no servers, no account, no telemetry
• Works offline after the one-time model download
• Optional upscaling (Real-ESRGAN) and face restoration (GFPGAN)
• Workspace gallery: browse your generations with full metadata
• Adjustable steps, guidance and seed; negative prompts; optional NSFW filter

Requirements
• Firefox 142+ with WebGPU available
• A GPU with enough memory for the model (≈ 8 GB recommended for SDXL). The extension checks your hardware on startup and tells you if WebGPU is off.

Privacy
Everything runs locally. No analytics, no remote inference. Generated images and all settings live in your browser's local storage on this device only.