Disappointing and misleading – not ready for real use.
This application does not deliver what it promises. Image generation is unnecessarily difficult and unreliable. Errors occur constantly, often without clear explanations, making it frustrating to get even a single usable result.
Human image generation is especially weak. Female characters all end up looking extremely similar, almost like sisters, regardless of how different the prompts are. There is very poor identity diversity or control, which completely defeats the purpose if you’re trying to create unique characters.
The new “canvas” feature is basically broken. Editing an image does not work as expected — instead of modifying the existing image, it frequently generates a completely different one, ignoring the original composition. This makes iterative work impossible.
Overall, the app feels unfinished and overpromised. The marketing suggests precision, control, and advanced editing, but in reality you get instability, errors, and inconsistent results. Not worth the time or money in its current state.
No support available on the site.