| Pick | Tool |
|---|---|
| 01 | |
| 02 | |
| 03 | |
| 04 | |
Start with the treatment, not the tool. Read down your shot list and sort every shot into one of four boxes.
The performance: the artist singing the lyric to camera. If you cannot film it, you need something that lip-syncs a person to the vocal across the whole song.
The inserts: the hands, the objects, the empty rooms, the impossible images. These are single shots, and you want control over each one.
The texture: a section, or a whole track, where the picture answers the music rather than tells a story.
The sketches: shots you want to test before anyone spends money on a location.
Each box has a tool that fits it best. Most videos use more than one, then meet in an editor such as DaVinci Resolve.
Muzie For the performance
One photo becomes the singer, lip-synced across your song.
One photo becomes the singer, lip-synced across the whole song, up to 6 minutes on a paid plan.
Reach for it when
- The treatment needs a face singing the lyric, lip-synced, across a whole song
- You have a photo of the artist but no day, crew or location
- You want a written visual direction (35mm film, claymation, anime) applied to every scene
- You need portrait 9:16 first and a widescreen cut for YouTube
Think twice when
- The concept is abstract, with nobody on screen: use Neural Frames
- You want to design each shot by hand: use Runway or Pika
- You already filmed footage and need to cut and grade it: use DaVinci Resolve
- Your song runs past 6 minutes
- Free plan
- Best for: A performance video when there is no shoot
- Web, iPhone, Android
Runway For insert shots
An AI studio for generating video, image and audio, one shot at a time.
One shot at a time from a choice of models, cut to the bar by you. 125 free credits once.
Reach for it when
- Your shot list has inserts you cannot film: a flooded ballroom, a field of red chairs
- You want to pick between several video models in one place
- You are happy to cut every clip to the bar yourself
Think twice when
- You want the whole song handled in one go
- You need a performer lip-syncing the lyric across the track
- You have no editor to assemble the clips
- Free plan
- Best for: Insert shots and impossible images, cut by hand
- Web, iOS, Android
Neural Frames For a visualizer
Audio-reactive AI animation and full music videos from your song.
Audio-reactive pictures that pulse with the track, and Autopilot for full-length videos.
Reach for it when
- The concept is a visualizer: the image breathes with the kick and the hats
- The track is long, and Autopilot is described for songs up to 10 minutes
- You want lyric or lip sync options on top
Think twice when
- You need a free plan to try it
- The treatment is a narrative with a realistic cast
- You need more than 25 fps
- No free plan
- Best for: Visualizers for club and instrumental tracks
- Web
Pika For sketches and remixes
AI video, image and audio studios, with swap, motion transfer and extend tools.
Rough out a shot before the shoot, or swap and extend footage you have. No watermark on any plan.
Reach for it when
- You want to test how a shot from the treatment might look before the shoot
- You have footage and want to swap an object or extend a clip
- You want no watermark, even while trying it
Think twice when
- You need a commercial licence on the free plan
- You want a full song cut for you
- You need a singer lip-synced across the track
- Free plan
- Best for: Sketching shots and remixing footage
- Web, iOS
Frequently asked questions
Which AI tool makes a whole music video from a song?
Muzie and Neural Frames' Autopilot both work from the full song. Runway and Pika make shots that you assemble in an editor.
Can AI make a music video look directed?
It looks directed when the brief is directed: framing, light, colour, movement and duration for each shot, written down before you generate anything.