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

Description and use

RapSong lets users create rap duet videos from two photos by automatically generating lyrics, beats, and lip‑synced face animation while preserving identity.

Problem

No simple workflow existed to let two people from static photos perform alternate rap lines while preserving facial identity, synchronizing lip movements, and generating custom beats and lyrics.

Outcome

RapSong enables casual users to create entertaining rap duet videos from two photos without camera shooting or video editing skills, addressing facial consistency, flickering, and lip sync challenges.

Workflow

• Upload two photos of your performers
• AI writes custom original rap verses
• Generate matching rap beats
• Animate both people trading bars inside the fixed orange studio backdrop
• Export vertical short clips ready for social media sharing

Evidence

Existing AI music tools can generate songs, and AI video tools can animate images, yet combining lyric writing, beat production, lip sync and character identity preservation in one single pipeline is tricky.

Source

The biggest technical challenge I faced is maintaining consistent facial features during animation, avoiding flickering, and synchronizing lip movements to the generated rap audio.

Source