Studio BTS Video Generator - Miniature Film Set Behind The Scenes
Studio BTS turns two picks - the place the miniature is of, and the disaster that hits it - into an 8-second video that looks like real behind-the-scenes footage of a movie shoot: a soundstage, a blue screen, an EFFECTS CREW, a full-size effects rig, and one practical effect going off on the model in front of the crew. The rig's noise and the effect's own sound are rendered with the video, not added afterwards.
What is a miniature film-set behind-the-scenes video?
It is a short vertical video shot like real BTS footage of a miniature movie set: a soundstage, a blue screen with tracking markers, a visible EFFECTS CREW, cinema cameras on tripods, and one impossible practical effect going off on the model in front of the crew.
The joke of the format is the reveal: what looks like a disaster movie is a waist-high model with people standing around it. Studio BTS generates the set, the crew and the effect together, so the miniature never has to exist.
Is the viral miniature wave video actually AI?
Yes. The wonderfulworldai Instagram reel of a giant wave hitting a miniature Miami passed 828M views on this exact format, and its caption initially framed it as a hand-built miniature and a 40-foot wave tank before disclosing it was AI generated.
Studio BTS produces that shot style from two picks. There is no model shop, no water tank, and no crew - the crew is part of the generated image.
How do I make one - what do I choose?
Two things: the place the miniature is of, and the disaster that hits it. Everything else - the soundstage, the crew, the cameras, the full-size rig, the camera move and the order the destruction happens in - is already in the template.
- •a coastal city, hit by a flood wave
- •a dense downtown of towers, hit by a tornado
- •a hillside town, hit by a wildfire
- •a harbour front, hit by a sandstorm
Do I need to upload a photo of a model?
No. Studio BTS needs no photo and no footage - it is text-to-video, which is the main thing that separates it from other behind-the-scenes AI tools: those animate a photo you supply, so you need a subject before you can start.
Because nothing is uploaded, the set can be anything you can describe rather than anything you can photograph - the place picker takes a custom answer as well as its presets.
Does it come with sound and effects built in?
Yes. Every render carries native audio - the hum of the stage, one shouted call, the machine spinning up, then the effect itself - generated in the same pass as the picture rather than added as a separate SFX step.
The order is part of the format: you hear the rig start before you see anything happen to the model. That is why the audio is written into the template's prompt rather than offered as a setting.
What format does it render in - is it ready to post?
An 8-second 9:16 clip at 720p, generated on seedance-2.5 in about three minutes - already framed for TikTok, Instagram Reels and YouTube Shorts.
The length is fixed at 8 seconds and the sound is always on - the beat spine and the effect's own audio are written into the prompt, so both are part of the format rather than settings. Aspect ratio and resolution are editable under Advanced, within what the model itself supports. The video lands in your Creations, where you can publish it to TikTok, YouTube and Instagram.
How much does one video cost?
9.6 tokens at 720p, and it is 9.6 at 720p every time - every Studio BTS clip is 8 seconds long, so the place and the disaster you pick never change the price. That is about 5 videos a month on Creator ($19.99, 50 tokens) or about 10 on Automate ($39.99, 100 tokens).
seedance-2.5 is priced per second, and 8 seconds at 720p is 9.6 tokens. Dropping to 480p under Advanced halves the rate, to 4.8. The generate button carries whichever number applies before you spend anything.

