Music Video Types and Export
Choose between the audio-reactive visualizer (with timed lyrics for a lyric video), a photo video built from generated scenes, a story video of AI clips with recurring characters and an AI singer video, then export the right aspect ratio and resolution.
Aimusicis makes four kinds of video, and the visualizer can also put the lyrics on screen. All four start from the same audio — a song from My Music, an upload or a recording — and differ in what you see.
Choose a video type
| Type | What you see | You provide |
|---|---|---|
| Music Visualizer | Graphics that move with the audio, over your image or a color | Audio; optionally an image |
| Photo Video | A sequence of generated pictures, one per scene | Audio and a visual direction |
| Story Video | AI video clips, one per scene, cut on the beat | Audio and a visual direction |
| AI Singer Video | A portrait that performs the song | Audio and a portrait |
Visualizer
Pick one of 12 templates, from Spectrum Bars to Starfield, among them Vinyl, a record turning with your cover art as its label, then add a title, an image and timed lyrics. The preview runs in your browser as you edit. Projects can be 16:9, 9:16 or 1:1 and export at up to 1080p. Audio files up to 50 MB and 15 minutes are accepted.
Lyric videos
A lyric video is a visualizer with timed lyrics. In the visualizer editor, a song generated here can bring its own timed lyrics in one step (Replace current lyrics with the selected song's timed lyrics); for an upload, use Auto captions or paste the lines and set their times. On the Lyrics tab each caption is a line with its start, end and text; fixing a misheard word keeps the line's word timing, and the timed LRC text sits under Edit as LRC text. Lyrics are shown as captions over the visuals.
Photo video
Describe the visual direction and generate scene prompts, review and edit each scene, then generate the pictures and the video. Scenes play in order, cut where the song's timed lyric lines fall, or evenly when it has none. Changing the audio, the scenes or the settings requires a new price quote before generating. Photo scenes are landscape or portrait.
Story video
Each scene becomes an AI video clip of up to 8 seconds made from its own picture, cut where the lyric lines fall and on the beat under your captions. Every clip is silent: the song is the soundtrack. Choose the Clip model; the price of one clip is:
| Clip model | 720p | 1080p |
|---|---|---|
| Veo 3.1 Lite | 27 | 43 |
| Hailuo 03 | 52 | 83 |
| Kling 3.0 | 90 | 115 |
Each scene also costs its picture. Choose how many recurring Characters the story has (none, or one to three) and how many recurring Locations (none, one or two): each is drawn once as a reference picture, and every scene it appears in is drawn from it, so a character keeps one face and a place keeps its look from shot to shot. Each character or location adds one picture to the price. With the scene editor you can rewrite the characters, the locations and the scenes, and tag who and where each scene shows, before generating. If a clip model refuses one scene's clip, that scene plays its picture and only that clip is refunded.
To review the pictures before paying for clips, choose Make: Storyboard first. The characters, locations and scene pictures are drawn and play under your captions as a storyboard video, at the price of the pictures and the render, with no clips.
The story is saved as a project. Open it from My videos (Open in editor) or from the projects list in the visualizer editor to move or remove a scene, or press Redo on one: rewrite it, then make a new picture and clip, a new clip from its picture, or a new picture only. Animate scenes makes a clip for every scene that still shows its picture, such as a storyboard's, in one run. Only those scenes are made again; the whole video renders again with every other scene kept. A redo costs each clip at its price and 9 credits for each new picture, plus 20 credits for rendering the video again.
AI singer video
Choose the song (or a section of it), then the performer: upload a clear portrait, pick one of your generated images, or generate a new portrait. A WebP picture is converted to JPEG before upload. Describe the performance and generate: the button shows the price for your selection. The song is performed in sections of up to 30 seconds made side by side and joined into one video. It costs 8 credits a second at 720p and 16 at 1080p, plus 20 for joining the video.
On a finished photo or story video, Make a frame sing opens the singer with the frame you pause on as the portrait and that video's song already chosen.
Before you export
- Aspect ratio: 16:9 for YouTube, 9:16 for Shorts, Reels and TikTok, 1:1 for feeds that crop to a square.
- Resolution: 1080p for upload to a platform; lower resolutions render faster for a draft.
- Cost: the visualizer preview is free; the export, scene planning and every generated video show their credit cost on the button before you start.
Finished videos are listed in My videos and download as MP4. Use these settings on one of them puts its choices back in its form, and a video that is still being made can email you when it finishes.
Related: Find, Share and Download Songs · Credits and Plans