I think of gemini have capacity to read multiple image's instead of one at a time it will great. For app prototyper. I think it requires improvement like currently it is suffering from invalid arguments. I think it requires more error handling. And also. The voice mode will Great and for app prototyper if there will option to add own api key for ai will more powerful.