Hi, I’m Jordan Hale.
I’m a 35-year-old commercial visual freelancer living south of Austin, Texas. For six years I’ve created product images, ad concepts, moodboards, and social assets for e-commerce brands, advertising agencies, and independent creators. In 2023 I started integrating Midjourney, Stable Diffusion, Flux, and ChatGPT Image into real client pipelines—not as magic replacements for creative judgment, but as tools that had to earn their place under deadline.
Prompt Receipt is the record of those tests.
Why this site exists
Most AI image content falls into two categories: flashy one-off generations that look great in a feed, or generic prompt lists that fall apart the moment a real commercial brief shows up. I wanted something different—a place that treats AI image tools the way a working freelancer actually uses them.
Every workflow on this site has been run against a defined job. You’ll see the brief, the prompts, the settings, the failure points, and the conclusion. If a method only works in a vacuum, it doesn’t make the cut. If a tool looks impressive in a demo but slows down delivery or increases revisions, I’ll say so.
The editorial promise is simple: no untested claims, no recycled prompt lists, and no vague AI futurism. Only workflows that survived a real commercial brief.
Who I am (and who I’m not)
I’m not an AI researcher.
I’m not a prompt-engineering influencer.
I’m not here to declare universal winners or predict the future of generative models.
I’m a working commercial freelancer who tests tools the same way I use them for clients: under constraints of time, consistency, brand requirements, and revision cycles. My credibility comes from documented briefs, controlled test batches, and the willingness to show what failed.
I keep a folder of failed generations. The bad ones usually explain the workflow better than the polished ones.
Life outside the studio
I’m married to a UX designer. We don’t have kids. We do have Biscuit, a rescue dog who has claimed a permanent spot near the desk. I ride bikes around Austin-area routes, make pour-over coffee at home, and regularly reorganize the garage studio that doubles as my production space.
Those details stay in the background. This site is about the work—how to get usable images out of AI tools when a client is waiting, not lifestyle content.
What you’ll find here
Client Brief Lab — realistic workflows built around commercial assignments: product shots, ad concepts, campaign directions, and social assets
Prompt Under Pressure — prompt construction for consistency, composition, product placement, brand style, and controlled variations
Model Bench — Midjourney, Flux, Stable Diffusion, ChatGPT Image, and related tools tested against defined commercial tasks
Failure Files — postmortems of rejected images, broken prompts, inconsistent characters, bad compositions, and expensive workflow mistakes
You’ll see exact prompts, settings, generation counts, time costs, and the moments where a tool or method broke. Conclusions are always tied to a specific use case.
My promise to you
Test it before you trust it.
Here’s what survived the brief.
The demo looked great. The deliverable did not.
I ran the test so you can skip the expensive part.
Useful is a higher bar than impressive.
The model wasn’t the only thing I got wrong.
I will never publish a tool recommendation I haven’t personally stress-tested. I will never hide the failures. I will never expose client identities, private briefs, or confidential pricing. And I will never treat “looks impressive” as the same thing as “is usable under deadline.”
If you’re looking for speculative AI futurism or one-click magic, this isn’t the place.
If you want tested workflows, honest tool comparisons, and the receipts that show what actually worked, you’re in the right studio.
Welcome to Prompt Receipt.
I’m glad you’re here.
— Jordan Hale