thinkhard.SI

Prompt repository Image & video

AI image and video prompts for work

8 scene descriptions for image and video models: a studio product photo, a professional headshot from your own photo, a blog header, a social graphic with readable text and short product videos. Each one names the lighting, framing and what must not change, which is what image models need.

Prompts
8
Registry code
MED
Average score
9.8 out of 10
Works in
ChatGPT, Claude, Gemini and Copilot

The 8 prompts

  1. MED-001Midjourney studio product photoYou need a clean, catalog-style product image for a store listing, an ad or a mockup.Midjourney9 out of 10
  2. MED-002Blog header illustration with a rendered headlineYou want a header or social image with a short headline set in it, from a plain-language image model such as Nano Banana or GPT Image.Image, plain10 out of 10
  3. MED-003Product video shot with sound, for Veo or RunwayYou need a short b-roll clip of a product for a landing page, an ad or a social post, from a text-to-video model.Video model9 out of 10
  4. MED-004Edit a product photo: new background, product unchangedYou have a real photo of a product and need it on a clean or styled background for a listing or a slide, without the product itself being altered.Image, plain10 out of 10
  5. MED-005Training video shot of a workplace taskYou need a short clip showing one step of a real task, such as lifting a box safely or scanning an item at a checkout counter, for a training module, from a text-to-video model.Video model10 out of 10
  6. MED-006Turn your own photo into a professional headshotYou need a clean headshot for LinkedIn or a staff page from a casual photo of yourself.Gemini10 out of 10
  7. MED-007Create a social media graphic with readable textYou need a post image with a short headline and your colors, sized for a specific platform.GPT10 out of 10
  8. MED-008Design an event flyer or posterYou need a printable flyer with the event name, date, place and a clear visual.GPT10 out of 10

How to use them

  1. Open a prompt and press Fill in variables. Type your details into each blank; they stay in your browser.
  2. Copy the filled prompt into ChatGPT, Claude, Gemini or Copilot, or check it first in the Prompt Grader.
  3. Read the answer against your own facts before you use it. The Fact Check List shows what to check.

Questions

How do I get a professional headshot from ChatGPT?

Upload a clear photo of yourself and describe the setting, clothing and lighting you want, and tell it to keep your face unchanged. The headshot prompt here spells all of that out, and works in ChatGPT and Gemini.

Do these prompts work in Claude, Gemini and Copilot?

Yes. Each prompt is laid out for one kind of model, shown beside it, but every one works in ChatGPT, Claude, Gemini and Microsoft Copilot. Fill in the blanks and paste the whole prompt into the chat.

Are these prompts free?

Yes. You can open ten prompts without signing up, and a free email unlocks all 319 in the repository. Use them in your own work and your clients' work; please do not republish the collection.