I've been testing Leonardo AI for a few weeks on my home server project. The image generation is solid, but I keep hitting walls with the canvas editor.
For instance, trying to extend a background or make a small tweak to one element feels clunky. Inpainting/outpainting controls aren't as responsive as in some other local tools I've tried. The layer handling also seems less intuitive. Is this just me, or have others found workarounds?
I'm hoping to use it for generating assets for a small game. Maybe I'm missing a setting or a different approach?
I'm running Leonardo's cloud plan at a 12-person marketing agency for quick content mockups and asset generation.
**Canvas tools vs. dedicated editors:** It's definitely a step behind. The outpainting is more for expanding a whole scene, not fine edits. I've had better luck regenerating the whole area at a higher strength than trying subtle tweaks.
**Inpainting latency:** On the Pro plan ($30-48/mo), I see about a 15-20 second turnaround per inpaint at standard resolution. It's fine for our batch work but feels slow for iterative edits.
**Workflow for assets:** For our game asset needs, we generate the base image in Leonardo, then pull it into a proper editor like GIMP or Photopea for layer work and precise adjustments. The in-platform editor just isn't there yet.
**Where it wins:** Speed of initial generation. For cranking out 30-50 concept variations in an hour, it's great. The editor is just a bonus for minor fixes, not a core feature.
For your game assets, I'd stick with Leonardo for the initial concepts because the generation is strong, but plan to use a secondary tool for the canvas work. If you need everything in one tool, look at a more UI-focused option.