About this tool
What is the Video Sequence Builder?
The Video Sequence Builder writes segmented prompts for AI video models — the multi-shot format that gets you a real edit inside one clip instead of a single drifting take.
You describe the beat in plain language (or speak it) and an agent writes the shot list: each segment gets its own timecode range, action, camera move, lens and sound, wrapped in the scene-level blocks the models actually weight — SCENE CONTEXT, ACTIVE REFERENCES, LOCATION MAP, PHYSICS, LIGHTING, STYLE and POSITIVE LOCKS.
Then you fix the timing, which is the part nobody wants to do by hand. Drag a shot's duration and every downstream (0.0s–1.6s) range rewrites itself, along with the segment and cut counts stated in FORMAT MODE. Lock a shot and "fit to 8s" scales the others around it. Time is tracked as whole frames at 24fps, so three 0.4s beats actually sum to what the total says.
It also warns you before you burn a render: past the target model's cut budget identity drifts, dialogue over 2.5 words per second desyncs lip-sync, two camera moves in one segment produce jitter, and asking a video prompt to draw a logo corrupts it. Pick your target model and the advisories follow.
This tool produces text. It never generates video, so there is nothing to pay for per render — copy the prompt into Seedance, Veo, Kling, Runway or whatever you use. The first 2 agent runs are on us; after that add a Google AI Studio key, which is free to create, and it becomes unlimited.
FAQ · VIDEO SEQUENCE BUILDER
Questions about the Video Sequence Builder
Does this generate video?
No, and that's deliberate. It produces a prompt you copy into whatever model you already use — Seedance, Veo, Kling, Runway, Sora. Because it never touches a video model there's no render cost, no queue and no watermark.
Why do I need my own API key after 2 runs?
The agent that writes the shot list runs on a large language model, which costs money per run. Everyone gets 2 free runs on ReelStack's key. After that you add a Google AI Studio key — free to create at aistudio.google.com, with a generous free tier — and the tool becomes unlimited. Your key, your quota, and we never see it in plaintext: it's encrypted before it's stored.
What's a segmented prompt and why does it matter?
Current video models can cut inside a single clip if you label the shots. An unlabelled multi-shot description renders as one continuous take; a prompt with SEGMENT headers and HARD CUT labels renders as an edit. Segmenting is also how you control pacing — a 10-frame flash cut next to a 4-second hold.
Can I save and share looks?
Yes. Save a prompt's STYLE block as a reusable look, along with camera moves and whole blocks, and reuse them across sequences. Anything you save can be shared with the community, and you can fork anyone else's shared look or sequence into your own library. Everything is private by default.
How accurate are the warnings about cut budgets?
They come from measured per-model limits: Seedance 2.5 handles about 6 shots in one clip, 2.0 about 4, Veo and Kling about 3. Past that, character identity typically drifts from the fourth cut onward. The tool warns and lets you proceed — it never blocks you.