Building RapSong: AI Tool to Turn Two Photos Into Hotel Lobby-Style Rap Duet Video

AI-powered video creation has seen rapid progress, but a new tool called RapSong stands out by combining several complex elements into one seamless process. The motivation behind this innovative SaaS project stems from the popular Hotel Lobby rap trend that went viral online. While existing AI music generators can create songs and video tools can animate images, fusing songwriting, beat production, lip-syncing, and preserving identities of subjects in one pipeline poses significant challenges.

Using RapSong is remarkably simple - users upload photos of two performers and the tool automatically crafts custom original rap verses for them. It then generates matching beats to perfectly sync with the lyrical content. Next comes the animated lip-syncing stage where both individuals perform their bars within a fixed orange studio backdrop. Finally, the system exports short vertical clips that are ready for instant sharing on social media platforms.

This entertaining AI video tool caters to casual entertainment needs - birthday roasts, wedding surprise clips, class reunion performances, corporate team shout-outs, and even humorous pet duets. No camera shooting or video editing skills are required to produce these fun results. The main technical hurdle faced by the creators was preserving consistent facial features during animation, avoiding flicker artifacts, and synchronizing lip movements with the generated rap audio. Continuous model tuning has been employed to enhance stability and visual quality.

To experience RapSong firsthand, interested individuals can visit the tool's website and create their own custom rap duet video from two photos. The creators eagerly welcome feedback and are curious about the types of imaginative AI-generated videos users would like to see next.