Role + task
Every prompt opens by telling the AI who to be and exactly what to make — no guessing.
Type what you want in plain words — or upload a reference image or video — an image, a piece of writing, a video, code, anything. Hit Generate prompt and get a complete, structured prompt you can paste into any AI tool.
Reference — optional
Upload a photo, design, screenshot, or artwork — or a short video clip. The generator reads its size, framing and colours on your device only (nothing is sent to a server) and builds them into the prompt.
Only one reference at a time — adding a video replaces an image, and adding an image replaces a video.
What should the prompt do with this image?
Builds a prompt to recreate the reference as closely as possible — same subject, framing and colours.
What should the prompt do with this video?
Builds a prompt to recreate the reference video as closely as possible — same subject, camera, pacing and colours.
Your idea
One sentence is enough. With an image or video uploaded, this becomes optional — use it for any extra direction, like “make it luxury” or “same scene at night.”
Templates
Framing — for image prompts
Video mode — what kind of video prompt?
Pick a mode and the structure below is baked into your generated prompt — shot by shot for ads, a full package for YouTube, native vertical for Instagram / TikTok.
Video structure — applies to every video prompt
Tool settings, not prompt text: duration, aspect ratio and resolution are set in the video tool itself — the prompt below names structure, camera, light and sound, and the generated prompt carries this rule so the words aren’t wasted on settings.
Commercial — built one prompt per shot
“No AI text” leaves clean empty space in frame for real text, prices and logos added later in editing. The end card is a designed shot: product centred and still, held a couple of seconds, room at the top for the call to action.
YouTube — a package, not just a clip prompt
Output: storyboard scenes (spoken line, visual type, duration, viewer purpose), per-scene B-roll prompts, chapters, end screen, and a designed thumbnail moment. No “hello and welcome” opener — the hook comes first.
Try an example
Your prompt
Open in
Each Open button copies your prompt first, then opens the tool in a new tab — just paste it there.
Copy formatted for
Standard — the full structured prompt, as generated.
Library
Every prompt you generate is kept here, on this device. Tap one to open it again; star the best ones to pin them to the top.
Every prompt opens by telling the AI who to be and exactly what to make — no guessing.
Subject, setting, style, and what to avoid are spelled out, so the first result lands close.
It states how the answer should come back — length, structure, and quality bar included.
Instagram / TikTok — native vertical 9:16
Composed for vertical from the start — never cropped from landscape. First frame is a verb hook: the action is already moving. One idea per video, one closing action.