Create original songs in one click

What Ribbi did

Ribbi unpacks "vintage vinyl, coffee shop, artsy clothing-store mood" into music style parameters, picks full-track mode, and generates two instrumental variants of about two minutes each (134.7s and 136.2s) in one go, with a full note on the generation settings so the user can listen first and choose. The user replies "the first one's great" — and the first creative line is locked.

What Ribbi did

Three consecutive rounds take the cover from "looks right" to "ready to ship": version one clones the reference with the title set to LO-FI JAZZ; the girl is swapped for the user's own 3D character and the two credit lines — Music by Ribbi, Cover by Ribbi — are added, re-rendered at 4K (5504x3072); the final round adds subtle facial freckles, realistic skin texture and Kodak film grain, landing a release-ready film-grain edition. Each round changes only what's pointed at — everything else stays locked.

What Ribbi did

Ribbi reads the image first — dark, mysterious, premium; technology colliding with humanity; futurist aesthetics — and translates it into a "rhythmic electronic + abstract atmosphere + Tech Noir" musical direction, keeping the abstraction deliberately restrained to protect the premium feel, then generates a five-minute and a four-minute instrumental. After listening, the user promotes the blue-iris image to cover art: same text layout as the Lo-fi Jazz cover, headline swapped to TECHNO — one album's visual language now reused across two creative lines.

What Ribbi did

On receiving the font file, Ribbi first states a real limitation: image generation models can't load an .otf file, so it approximates Trap Black by writing its traits — ultra-bold, geometric, high-contrast uppercase — into the visual description. Then it renders by the Lo-fi Jazz layout rules: a golden-orange POP headline over the blue-iris key visual, both credit lines kept in place, 4K resolution. Three covers side by side — a series typography system has taken shape.

What Ribbi did

The first burn smoothly delivers a 56.7-second cover video, but transcription finds exactly one word of "lyrics" — an "Oh": instrumentals have no words to subtitle. Ribbi lays out the choice: keep the instrumental and drop subtitles, or regenerate with vocals. The user picks vocals and sets a cyberpunk direction; two vocal electronic tracks arrive within minutes. With a version chosen, the first 50 seconds are re-burned and the subtitle animation lands on a precise three-line lyric timeline — white glowing outline, 0.3s fades — closing out from a one-sentence request to a subtitled music video.