Has anyone here used image-to-image tools for early streetscape or transit concepts? For Jacksonville projects, a useful workflow is to start with a real street photo or a simple massing sketch, make a few versions with the same prompt, and then compare lighting, tree canopy, storefront rhythm, and pedestrian space. I think it works best as a visual conversation starter rather than a finished plan—label what is real, keep the original beside each variation, and verify dimensions, drainage, and accessibility separately.
For anyone who wants to try that workflow, this image to image ai (https://imgtoimg-ai.com/) tool is a straightforward place to test image generation/editing from a reference. It could be useful for exploring facade color, streetscape planting, or transit-stop ideas before a designer turns a concept into something buildable. What safeguards or local data sources do others use when sharing AI-assisted visuals?