Contents
A music video is easy to imagine and hard to finish: you need the track itself, dozens of shots, the same hero throughout and an edit that fits the music. AI takes care of the shots and the sound, but the order and the rhythm stay with you.
The short answer: make or upload a song, split it into scenes by structure, generate a photo or a short video for each scene, then assemble everything on the Montage timeline over the music track and export to MP4. Close-ups of singing can be brought to life with lip-sync models.
Givon AI does not make a complete music video from an uploaded song with one button. But the whole route lives in one project: track, scenes, takes and export, with no moving files between services.
What you need
- A song: your own recording, or an idea to generate one (genre, mood, lyrics).
- A hero look or a plot: what happens in the verses and in the chorus.
- A few dozen tokens: a music video takes many generations, and the cost of each is shown before you run it.
Step by step
- 1
Prepare the song
Generate a track in music mode, or upload your own file in the Music tab in Montage. A short excerpt works best for a music video: a verse and a chorus, 30–60 seconds.
- 2
Split the song into scenes
Listen to the track and write down the blocks: intro, verse, chorus, bridge. Give each one a single shot idea. Models make clips of 5–15 seconds, so a 20-second chorus is 2–3 scenes.
- 3
Lock in the hero
If the video is about a character, first build the look as a reference and use it in every scene. How to do it is described in the references guide.
- 4
Generate the scenes
Get a photo or a video for each block of the song. Start with photos and animate the good frames: finding a composition is cheaper that way than trying out videos one by one. Details are in the image-to-video guide.
- 5
Bring the vocals to life
For close-ups of singing, build a talking shot with a lip-sync model: a photo or video of the hero and an excerpt of the track.
- 6
Assemble the video on the timeline
Put the scenes in order, place the track on the music lane, trim the edges of the scenes to match the changes between blocks and watch the whole video.
- 7
Choose a format and export
Use 9:16 for Reels and Shorts, 16:9 for YouTube, 1:1 or 4:5 for the feed. Export creates an MP4.
The song: generate it or upload your own
There are two routes. If the track is already recorded, upload it: the Music tab in Montage, the Import file button. The rights to the composition and the recording stay with you, so upload only what you are allowed to use.
If you have no track, describe one in music mode. In Suno you set the style, the lyrics and the voice: Auto, female or male, and for background music you can turn on instrumental. You get two versions to choose from. The model picks the length of the track itself, so for a music video you will most often need to trim it on the timeline.
Energetic indie pop with female vocals, guitars and synths, warm vintage sound, 110 bpm, bright atmosphere, a catchy chorus
In the style, describe the genre, mood, tempo and instruments. Paste the lyrics separately into the Lyrics field, or leave it empty and the model will write them itself.
- Suno V6Finished songs with your own lyrics, or instrumental music for video: two versions to choose from.from6 tk
- Suno V6 MiniA lighter version for quick drafts: check the lyrics and try out genres before the final recording.from6 tk
- MiniMax Music 3A finished WAV track from a description and required song lyrics, with verse, chorus and bridge tags, up to 5 minutes.0.2 tk/ s
Scenes that follow the song structure
Do not try to get the whole music video from one prompt. A music video is built from scenes: each covers its own piece of the song, while the hero, the clothes and the style hold on a reference. For a plot with a single hero, the models that keep the look between scenes work best.
- Kling 3.0Up to 15 seconds per scene and a character look that holds between scenes, suited to story-driven music videos with a single hero.from3 tk/ s
- Seedance 2.5Scenes up to 30 seconds from text, frames and references: enough for a whole chorus in one generation.from7.5 tk/ s
Weak
A beautiful scene for the chorus, dynamic and emotional.
Better
A girl in a red jacket dances on a rooftop in a night city, the camera slowly circles around her, neon signs in the blurred background, wind in her hair, warm light, cinematic style.
One main action and one camera move per scene. The track sets the tempo and the mood, so the prompt does not need them.
If the hero has to sing on camera
Lip-sync models match lip movement to an audio track: you give them a photo or a video of the hero and an excerpt of the song. The catalog describes them for speech, and we have not tested them on singing separately. Try a short excerpt first, and only then do a whole verse.
- Sync Lipsync v3Matches the lips to the track on a photo or a finished video of the hero, track up to 5 minutes.8 tk/ s
- H3 Max Lip SyncOne photo plus finished audio: the clip runs as long as the audio, up to 14.8 seconds, from 480P to 2K. For short lines and phrases.from3 tk/ s
- Kling Lip-SyncRe-maps the lips in a finished video to a new track, track from 2 to 60 seconds.3 tk/ s
Assembling on the timeline
In Montage the scenes run in one sequence with hard cuts. You can move, trim, split and replace them, and the track sits on a separate music lane. Replace a weak scene without rebuilding the video: everything else stays in place.
You will have to match the cuts to the beat by eye: there is no automatic rhythm sync. For that, it is best to cut scenes at the boundaries of the song blocks: verse, chorus, bridge.
Video from photos to music
The simplest music video is a slideshow of photos set to a track. Add the photos to the timeline as scenes, set their duration, put the music under them and export. To make the frames move, animate them one by one: image to video turns a photo into a clip of a few seconds.
Such a video is faster and cheaper than a full one: photos cost less than videos, and the rhythm comes from the duration of the scenes.
Limits and what to expect
A music video is many generations
A one-minute video is 6–10 scenes, each with several takes. Start with the chorus and make a short version for Reels.
Editing without effects
No keyframes, transition effects or multi-layer lanes: only scenes, sound and captions. For complex effects you will have to finish the video in another program.
The track and rights
Upload your own track only if you hold the rights to it. Check a generated track against the rules of the platform where you publish.
FAQ
Can I make a music video for free?
You can try it: after you sign up, 30 tokens arrive on your balance and no card is needed. They will most likely not be enough for a whole music video: it takes many generations, and the price of each is shown before you run it.
Can I upload my own song?
Yes. In the Music tab in Montage, click Import file and choose the audio. After that the track goes onto the music lane like any other.
Which frame and export formats are available?
9:16, 16:9, 1:1 and 4:5. Export creates an MP4 at 30 frames per second.
Will the video follow the rhythm of the song on its own?
No. You place the cuts by hand on the timeline: there is no automatic beat sync.
Will the model sing in my voice?
In Suno you choose the voice from three options: Auto, female or male. If you need your own vocals, record the song yourself and upload the finished file.
