Everyone else starts you at a blank prompt box. Prompt Reverse starts from something that already works.
Generating an image from a prompt is the easy part, and every AI tool already does it. The hard part is knowing what to ask for: describing a look precisely enough to get it back is a skill, and it is the skill nobody wants to learn in order to post an ad.
So we inverted it. You can already see which ad is outperforming yours, which listing photo ranks above yours, which thumbnail owns your niche. What you could not do was turn any of it into a brief. The reverse does that: it reads a reference and writes down what makes it work, so you start from a proven result instead of a blank field. The prompt is the thing you keep, which is why the second thing you make here is faster than the first and the tenth still matches.
Pay for what you use. One-time credit packs, no subscription, nothing recurring to cancel. Credits never expire, so a quiet month costs nothing, and the same balance covers the web app, the API and the MCP connector.
Show the cost before you spend it. Every tool and guided kit shows what it will cost before you start, and a generation that fails is refunded per image.
Your saved work stays yours. The prompt library is read-only over both the API and the connector, so an integration with a bug can never rewrite work you have kept.
We also try to be straight about what AI is still bad at. It drifts, it occasionally ignores you, and it is not a replacement for a designer on work that has to be exactly right.
110 tools across image, video and audio, guided kits that chain them, and an API and MCP connector over the same loop. The fastest way to understand any of it is to point it at an image or a sound you like and read what comes back.