AI Music Video Generator from a selfie

Straight answer first. Each render is one short clip, 5 to 15 seconds long: the people from your photos, framed from the waist up, each performing into their own vintage microphone on a floor stand. It's a single locked-off shot with no cuts, no zoom and no scene changes. It plays like a short live-session clip. A three-minute music video with a storyline is outside what it does.

• Length: 5, 8, 10, 12 or 15 seconds (12 by default)
• Format: MP4, vertical 9:16 by default, or 1:1 or 16:9
• Quality: 480p, 720p or 1080p
• Camera: one static shot, waist-up framing
• Sound: AI vocals and an original beat, lip-synced to the faces
• Watermark: none

That format suits Reels, TikTok and Shorts, where 12 seconds is plenty to land a hook.

Many music video tools start from a track you already have. Kiravu starts from a photo and makes the song as well. You pick a beat style and write the words. Kiravu produces the beat and the AI vocals, then lip-syncs them to the faces in your photos.

Under the hood it runs on ByteDance Seedance, which builds the video from reference images and generates the audio with it. Your photos tell it who performs. The stage you pick is the set. Your lyrics and beat tell it what to perform. Nothing is filmed, and nothing is borrowed: Kiravu never uses an existing song's recording or lyrics.

The words are yours to shape. In a 12-second song you get two verses of about 16 to 22 words and a closing hook of one to six words; shorter clips get shorter verses. In a duet, the first person takes verse 1, the second takes verse 2, and both take the hook. In Solo mode, one person performs all of it.

Some tools need a finished track, some stay silent, some write the song. Our comparison of AI rap video generators shows which is which.

There are 9 stages and 3 beats. This page starts on Rooftop at night with the Melodic beat: warm string lights, a blurred city skyline and a sung pop and R&B track at about 100 BPM.

• Love songs and slow jams: Rooftop at night or Marble lobby, with Melodic.
• Hype and flexing: Neon studio, with Trap.
• The viral orange-booth look: Hotel Lobby orange booth, with Trap.
• 90s rap energy: Graffiti alley, with Boom bap.
• Parties and roasts: House party or Battle stage.
• Holidays: Christmas lounge or Halloween night.

Melodic means the performers sing. Trap and Boom bap mean they rap.

Short is the point. A clip this length loops on a feed, fits in a text and is easy to watch to the end.

Post a duet with your partner as a Reel. Send a birthday or anniversary song instead of a card. Make a hype clip of yourself for your profile. Celebrate a team win or a graduation with the friends who were there.

Want something longer? Render a few clips on different stages and join them in any video editor. Each clip stands on its own, so the cut is up to you.