Turns "make one like this for us" into a fixed process. It fetches the reference video and measures how many shots it has, how long each runs, how the on-screen text is set, and how it opens and closes, then writes a breakdown. It rewrites that structure into your own shot list, on-screen copy and voiceover from your brand profile, tells you how many shots it will generate, and waits. After approval it generates shot by shot, checks every clip, cuts to the same rhythm, adds on-screen text, voiceover and music, and delivers the film, a cover and the post copy. It remakes structure and rhythm. The reference's footage, music and people never enter your film.
X links are fetched directly; for Douyin, Xiaohongshu, Channels and others you upload the file. Shot count, median shot length and pacing by section are measured, on-screen text is transcribed, and a breakdown is written.
The hook, section order and cut rhythm stay; the content becomes your brand, product and selling points. Every shot lists picture, duration, on-screen text and voiceover.
The number of shots and total seconds to generate are listed and approved first. Characters and products are locked with reference stills before any video is made.
Frames from each generated clip are inspected for limbs, faces, props and stray text. A failed clip is regenerated once; if it fails again a still replaces it and is flagged.
Shots are trimmed to the measured lengths. On-screen text is added in post with real fonts, voiceover lines land on their shots, and music ducks under the voice.
The film, a cover image, a title and a description, with a list of which shots are stills and what you should listen to before publishing.
Plain language works. These are real requests.
Make a 30-second version of this for our product, same structure and rhythm: <link>. Breakdown and shot list first, don't generate yet.
I uploaded a Douyin hit. Break down why it works: shot count, pacing, caption style, open and close.
Shot list approved, start generating. Use the character from the asset library and an English female voiceover.
X (Twitter) links are fetched directly. For Douyin, Xiaohongshu, WeChat Channels, YouTube and TikTok you currently download the video and upload the file; the breakdown and remake are the same from there.
It remakes structure, rhythm and layout, which are common editing techniques. The reference's footage, music, people and copy never enter the film; all content is replaced with yours.
If you provide a music file, it uses yours. If not, it synthesizes an instrumental track at the reference's tempo and section structure, with no samples and no third-party music, so it is clear for commercial use. The reference's own track is never used.
It analyses the picture and on-screen text only; there is no speech transcription yet. Tell it whether the reference has a voiceover and what it says.
It can be enabled on the Starter plan. Usage is deducted from your credits and goes mostly to seconds of generated video; breakdown, editing and music are not charged separately. Before generating, it quotes how many clips and how many seconds, and starts only after you approve. When credits run out you can buy an add-on or upgrade.
Creative Production makes images and video in batches from a brief. Viral Remake does one thing: you give it a reference video and get back a finished film for your brand. They are enabled separately.
Free credits on sign-up. For enterprise delivery, book a walkthrough and we will map your process to a job description.
Free credits on signup · No credit card · Mac and Windows