From a finished track (made in Suno or anywhere else) to a cut, on-the-beat music video with the same characters throughout.
You have a finished song. Maybe you made it in Suno, maybe you recorded it. Now you want a music video where the singer and the people around them look the same from the first shot to the last, and the cuts land on the beat.
Here is the order that works, step by step, as Inkreel does it.
Before any pictures, write down who and where this video is about: the artist or character, the people around them, the places, the look. In Inkreel that's the studio's IP bible, a plain text file with a template you can download from Help. It's what every later step reads, so nothing about your characters has to be invented along the way.
Upload the song file and paste the lyrics. Inkreel finds the tempo, the beats and the sections (verse, chorus, bridge) on its own, for free. You can rename a section or move where it starts by hand. If you want the words timed, lyric timing reads where each line is sung; it's a small paid step and shows its price first.
There's no fixed number of scenes. A three-and-a-half-minute song is often about 30 scenes at around six seconds each, but slower songs want fewer and faster songs more. Inkreel suggests a range from the song's length.
The storyboard writer drafts the scene list from the IP bible, the lyrics and the beat map: what happens in each scene, who's in it, where, the framing, and how it moves. It may only use characters and places you've given it, so it can't invent a new band member.
Then edit it. Change any column, reorder, add, split, merge or delete scenes, and mark the scenes where the singer is seen singing. Each scene's start and end snap to the beat. Approve the storyboard when it reads right.
Make a reference sheet for every character and recurring place in Cast, plus one style frame for the look. This is what keeps faces the same. See How to keep AI characters consistent between shots.
Make a still for each scene. Every still is made with the sheets of the characters and place in it attached. Look at them all, redo the ones that are off (with a note, or by marking up just the wrong part), and approve them. Stills cost a fraction of what video does, so this is where to spend your redos.
Each approved still becomes a short clip. In Manual mode you accept or redo each clip before it goes on the timeline; in Auto mode you confirm the total cost once and everything runs. Scenes marked singing can be lip-synced to that stretch of the vocal.
A clip that's nearly right can be fixed rather than remade: say what to change, or pause it on a frame and mark the spot, and a clip-fix model changes only that.
The editor lays the clips on the timeline against the song. Trim, reorder, swap takes, add transitions and text, and ask the chat assistant for changes in plain words. Export a 16:9 video at 720p or 1080p with the full song as its soundtrack, and a 9:16 version for vertical platforms.
You see the price before every paid step, and each project shows what it has cost so far, per finished minute. Bringing your own key means you pay the AI provider directly at their prices. The beat map, editing and export are free.
