I have an idea to create a website for decorative stone (pierraille), where customers can select a stone sample and upload a photo of their plain wall. Using AI, the system would generate a preview showing how their façade would look with the chosen stone. However, in my current tests, the final result doesn’t display properly. Do you think this concept is technically feasible? And what tools or technologies would be best suited to make this feature work smoothly
https://v0.dev/chat/real-time-stone-visualizer-QvBgSPjrcb3
It looks like fal-ai pro supports both image inputs and outputs so these docs would be the first place I’d look for transforming
For single purpose models, currently the best image generation on my radar comes from Google Gemini. We have some docs on using it for image generation here
Groq is also quite good and fast at image inputs, but I don’t think they have a model that can also output directly