Simple AIWhat it does: Writes Midjourney prompts that use one of your pictures as a style reference, so every new picture looks like it came from the same shoot.
Use it when: Your website and posts mix different picture styles.
You get: A base prompt, the style reference setup, one ready prompt for each scene on your list (up to 5), and a check list.
The prompt
You are my picture prompt writer, a senior art director with 15 years of keeping a brand's photos looking like one shoot. I use Midjourney. Goal: write Midjourney prompts that use my picture as a style reference, so I end with new pictures that look like they came from the same shoot as my old ones. Context: I use these pictures on my website and posts. My customers want a brand that looks put together, because a mix of styles makes a small business look careless. First, check that you can read the picture I gave you (the file or the link). Tell me in one line what you find in it: the light, the main colors and the mood. If it's missing or you can't open it, stop and ask me to paste it instead, or to describe it in a few lines. Ask me these, then run. Skip any I already answered. 1. Which scenes do you need next, up to 5? (Default: if I don't say, write one scene marked (suggested) and add [check: my scenes].) 2. Where do they go? (Default: square posts, so --ar 1:1.) 3. What do you like about the picture? (Default: use the light, colors and mood you found in it.) The picture I like (link): [image link] Do this: 1. Write a base prompt: the look in words (light, colors, camera, mood) in under 30 words. Use only what is in my picture or what I told you. 2. Show how to add my picture as a style reference with --sref and the link, and the right --ar for where it goes. 3. Write one full prompt per scene, each using the base prompt and the same --sref. Only the scene part changes. The light, colors and camera words stay the same in every prompt, so each picture is made the same way every time. 4. Check each scene prompt against the base prompt word for word. If a look word is changed or missing, fix it. 5. Tell me to save the style code (or the link) so the next batch matches. 6. A check list for each result: same light, same colors, no text, no extra fingers, nothing that looks like a real brand. Example of one scene prompt, match this shape: "a baker sets fresh bread on a wood counter, soft window light, warm cream and brown tones, 35mm photo, calm morning mood --sref [your link] --ar 1:1" Give me: the base prompt, the --sref setup in 2 lines, one prompt per scene as a list, then the check list. Under 400 words in all. Before you show me, check: - The number of scene prompts equals the number of scenes I gave you. - Every scene prompt has the same base prompt words and the same --sref link. - The --ar fits where each picture goes. Rules: - Drafts only. You write the prompts. I run them in Midjourney myself. Don't post or share anything. - No famous people or brand logos. - If you have to pick something I didn't give you (a name, a time, a setting), mark it (suggested). - The example is only a shape. Take every look word (light, colors, camera, mood) from my picture, not from the example. - Use only what is in my data and what I tell you. Never make up a name, number, date, price, link or promise. - If something you need is missing, write [check: what is missing] in that spot and keep going. Last, save these steps as a skill called 'Match My Pictures' so you do it this way each time I make new pictures for my brand. To run it again, just say: "Run my Match My Pictures skill."
You need
Midjourney
Any AITested on sample data
Built on The E-Myth Revisited's rule to do it the same way every time.
Copilot version
There is no separate Copilot version of this one yet. Paste the prompt above into Microsoft 365 Copilot Chat.
See a real example
Tested on sample data Oct 29 2026, after the rewrite: a Claude agent ran it on made-up sample input (tool results simulated). A separate Claude agent judged the run and the card.
Watch: no video for this one yet.
AuthorCorey Leon
