Finally, I have to give a shoutout to @Github Copilot – it saved me a lot of time on this! 🤖

Node-based UI, supports text-to-image & multiple image-to-image (perfect for consistent characters). It also does inpainting (but with gpt-image-1 results may vary 😜). All browser-based, no server. Files stay on your device. You'll need your own OpenAI key.

Yeah...