AI image tools

This is the roadmap, stated plainly. None of the AI tools listed here are implemented: nothing is processed, no image is uploaded, and no result is produced by any of them. Their pages exist to document the intended workflow and to keep the architecture honest about what does and does not exist.

Two tiers are planned. Local AI runs inference on your own device with ONNX Runtime Web, using WebGPU where the hardware allows and WebAssembly where it does not — background removal and upscaling are the first candidates. Cloud AI covers the generative work no browser can do, and stays disabled unless an AI provider is configured on the server.

If you need something that works today, the image tools hub is where to go.