Skip to content

Script to Image Prompts

Paste your full script and get an image prompt for every stretch of it: the B-roll you still have to shoot or generate, in the order the script runs, plus three thumbnail prompts. The prompts are plain English, so the same list works in ChatGPT, Midjourney, Flux or any other image model.

0 words. At least 100 words are needed for scenes worth prompting.

Used to keep the prompts on topic. The script decides the scenes.

0 words — at least 100 needed.

Prompts done? Check the script itself, free

Paste the same draft. You get hook, structure and pacing scores and the one issue costing it the most, quoted from your own lines.

Free · No login · See a sample audit first if you prefer.

What you get

Scene-by-scene prompts

One prompt per 60 to 120 words of script, in script order, each quoting the lines it covers so you can match prompt to moment.

B-roll shot list

Every prompt names the subject, the camera angle, the lighting and the style, which is the part a one-line idea usually leaves out.

Three thumbnail prompts

Three thumbnail concepts with a prompt and the text overlay each one was written to carry, kept consistent with the scene list.

What this is not

These are suggestions, not images, and they are not a measurement of anything. This tool does not render a frame, does not promise a click-through rate, a view count or a retention curve, and cannot tell you whether a thumbnail will perform. It writes the prompts; the model you paste them into produces the picture, and you decide what ships.

How to use the prompts

  1. 1. Pick your tool

    Paste a prompt into ChatGPT or DALL·E, Midjourney, Flux, Stable Diffusion or Ideogram. Nothing in the wording is specific to one of them.

  2. 2. Keep the style note as a prefix

    The style note describes the look the whole list was written for. Add it in front of each prompt and the scenes come back looking like one video instead of 30 unrelated images.

  3. 3. Change what misses

    Edit the words, not the whole prompt. Swapping the subject or an angle is usually enough, and re-running one scene is cheaper than re-running the list.

  4. 4. Check the rights

    Every model has its own terms on commercial use. Read them before the output goes on a monetised video, and keep a record of what you generated.

Frequently Asked Questions

Which image tools do these prompts work with?

Any of them. The prompts are plain English with camera, lighting and style described in words, with no artist names and no tool-specific syntax, so the same prompt goes into ChatGPT (DALL·E), Midjourney, Flux, Stable Diffusion or Ideogram. Midjourney accepts the sentence as-is; some tools return a better frame if you add their own aspect-ratio flag, which you set in that tool rather than here.

Are these prompts safe to use on a monetised channel?

The prompts avoid naming living artists, real people, brands and copyrighted characters, which removes the most common cause of a refused or unusable generation. They cannot decide the rights question for you: every image model has its own terms on commercial use, and you are responsible for what you generate and publish. Read the terms of the model you use before you use its output on a monetised video.

How many scenes does it produce?

One scene per roughly 60 to 120 words of script, capped at 40 scenes, so a 2,000-word script comes back with about 17 to 33 prompts and a 10-minute script fills the cap. Each scene quotes the script words it covers, so you can check the prompt against the line it was written for. Three thumbnail prompts ship alongside the scene list.

Is my script stored?

Yes, briefly and privately. Your script is stored with the result so the tool can serve the same prompts back to you at your result link, and it is never published or shown to another visitor. The prompts are suggestions written by a model from your words, so read them before you generate from them.

Can it generate the images for me?

Not yet. This tool writes the prompts and stops there, and the card under your results is how you tell us if you want the rendering step too. No paid image pack exists today.

See where your next script leaks viewers

Paste your script, get your scores and the biggest leak for free. No login.