How to Create a Music Video with AI
A step-by-step MVFlow guide to music videos and photo-to-singing music videos, with a 15-credit storyboard preview, real examples, and screenshots.
To create a music video with MVFlow, upload your song, set a creative direction, generate an editable storyboard, render the shots you choose, and compose them with your original audio. For a singing music video, choose a Singing Video mode and upload a photo and vocal audio instead.
Try it. Sign in to save projects and use your credits. Eligible users receive free credits when they log in each day.
Choose your music video workflow
| What you want to make | Mode | What you need | What you create |
|---|---|---|---|
| A music video with a story, several scenes, and edited shots | MV | A song, creative direction, and optional reference images | A timed storyboard, shot versions, and a composed music video |
| A photo of a character singing your audio | Singing Video · Singer Control | A photo and vocal audio; an optional mask to choose who sings | A singing performance, with audio up to 10 minutes |
| A singing character with prompt-guided expressions and movement | Singing Video · Expression Control | A photo, vocal audio, and a performance prompt | A singing performance, with audio up to 5 minutes |
The MV workflow gives you control of a complete visual story. Singing Video is the direct route when you want to animate a performer from one photo.
Watch the music video tutorial
This is the same tutorial shown on the homepage. It demonstrates the MV workflow and includes a finished music video excerpt. Use the player controls to turn on subtitles in your language.
What to prepare
- Music: For MV, use a track between 4 and 600 seconds. For a singing music video, use clear vocal audio within the selected mode's duration limit.
- Images: Reference images are optional for MV. Singing Video requires a character image. A clear, visible face helps communicate the intended performer.
- Direction: Decide on the mood, genre, and visual style, or explicitly let AI decide. You do not need to write every shot prompt yourself.
- Format: Choose an aspect ratio and resolution for your release. Available settings depend on the selected mode.
- Credits: A complete MV storyboard costs 15 credits. Video generation and final composition are later steps; review the credit estimate before confirming each operation.
Use music, images, and likenesses you have permission to use.
How to create a full music video
1. Start with your sound
Start creating, sign in, and choose MV. Upload your track in Music source, or choose another available audio source in the workspace. Add a project title so you can find the project again.
The MV workflow accepts songs from 4 seconds to 10 minutes. The original track is used for the final composition.
2. Set your creative direction
Choose a music genre, an emotional tone, and a visual style. You can select a primary style and optionally a secondary style, or choose Let AI decide. Use Additional ideas to describe the story, setting, characters, and key moments you want.
For example: “A hopeful indie-pop song. A singer travels through a rainy city at night; the chorus opens onto a warm sunrise. Cinematic live action, gentle camera movement, recurring red jacket.”
Optionally upload up to five reference images for characters, locations, wardrobe, props, or visual style. Without uploaded references, AI designs visual resources from your direction. Check the aspect ratio, resolution, and video model in Settings before continuing.
3. Shape the storyboard
Select Generate storyboard. For 15 credits, MVFlow analyzes your music and builds the full storyboard, including timed shots, all image and video prompts, and automatically generated reference images. This is a preview of the production plan; the video shots have not been rendered yet.
Review the musical sections and the shots assigned to them. Check how the story develops through verses, choruses, and transitions before spending credits on video generation.
Tutorial still: the music timeline connects musical sections to the planned visual story.
4. Refine shots and shared references
Review the shared characters, locations, and props. Reuse references across shots to guide a consistent visual direction, and choose which references matter for each scene.
Edit shot timing, action, image prompts, and video prompts as needed. Give the camera a clear instruction and check each scene's references. Review any credit estimate before generating additional assets or regenerating the storyboard.
Tutorial still: shared visual resources can be reused across individual shots.
5. Generate and choose your shots
Generate one shot to explore a direction, or generate the shots that are ready. Review the credit estimate before confirming. Video rendering uses credits separately from the 15-credit storyboard.
Preview each result, compare available versions, and select the take you want for the final edit. If one scene needs another pass, refine that shot's prompt or references and generate another version. Select your preferred version for every shot before composing.
Tutorial still: review generated takes and choose the versions for your final cut.
6. Compose the final cut
When your selected shots are ready, use the composition action to assemble them with your original soundtrack. Review any displayed credit estimate, wait for the task to finish, and preview the result.
Check the full video for timing, visual continuity, and audio before downloading. If a shot needs changing, return to that scene, choose or generate another version, and compose again.
7. Download and keep your work
Download your finished MV and any generated assets you want to retain. The workspace stores generated resources for 7 days, so save your files promptly rather than relying on the task history as permanent storage.
How to create a singing music video from a photo
Use this workflow when you want the person or character in a photo to perform your vocal track. This example is also shown on the homepage; press play to hear the result.
1. Choose a Singing Video mode
Try it, then choose Singing Video · Singer Control to choose who sings, including an optional mask for images with more than one subject. This mode supports audio up to 10 minutes.
Choose Singing Video · Expression Control when you want to guide facial expressions and movement with a prompt. This mode supports audio up to 5 minutes. Check the active mode's input limits before uploading.
2. Upload your character image
Upload one image, generate an image using the available image tools, or choose an existing creation. Use a clear face and a composition that suits the performance you want. You can explore the built-in sample photos to understand suitable framing.
Actual workspace: select a Singing Video mode, then add a character image and vocal audio. The interface is shown in English.
3. Add vocal audio
Upload your vocal track or use another audio source offered by the selected mode. Listen to it first and choose the section you want to animate. For a singing performance, use audio that contains the intended singing voice.
4. Set the performance and review the cost
For Singer Control, use the optional singer mask when you need to identify the intended performer. For Expression Control, describe the expression and movement you want, such as “sing softly, with a warm smile and gentle head movement.”
Choose a resolution from the options shown for that mode. Review the credit cost and your available balance before selecting Generate. Singing music video generation is a separate workflow from the 15-credit MV storyboard.
5. Preview and download
Follow the generation task in the workspace. When it finishes, play the video with sound to check the performance, mouth movement, and framing. Download the result promptly; generated resources are stored for 7 days.
Try it, then choose a Singing Video mode.
Free credits and storyboard previews
Does MVFlow offer free credits?
Yes. Eligible users receive free credits when they log in each day. Your account shows your available balance and credit validity. You can use that balance toward the workflow, including the 15-credit storyboard. Daily credits do not guarantee that a complete rendered music video will be free.
Can I preview a music video storyboard before paying?
Yes. A complete storyboard costs 15 credits and includes all shot plans, image and video prompts, and automatically generated reference images. If your available free login credits cover 15 credits, you can generate it without buying credits. Video rendering and final composition are separate steps with their own credit estimates.
Tips for a better first video
- Start with a clear visual idea and a recurring character or location.
- Review the storyboard before rendering; adjust the plan while changes are easier to make.
- Try one video shot first to check the direction before generating the rest.
- For singing music videos, choose a clear face and clean vocal audio, then check the active mode's duration limit.
- Download your results within the storage period.
Pricing: compare credit options when you need more balance.
Ready to turn your song into visuals? Start creating.