Slogans.ai
Type a topic and get ten AI slogans, each rendered as four t-shirt designs.
You type a subject — cats, cycling, accountancy — and the app asks GPT for ten short slogans on that theme. For each slogan it also asks for a single English word describing a matching picture, so "a purrrrfectly happy animal" comes back tagged with "cats". Slogan and subject are then handed to a rendering API, which finds an image, composes the words over it and sends back four variations. The page fills in as they arrive: ten slogans down the screen, four mockups across each one.
It exists because I had already built the rendering half. Free Merch Maker is an image-composition engine of mine with an HTTP API, and I wanted to know whether a language model could do the part in front of it — the thinking up of what to say and what to show.
There is no render queue. Each composition is base64-encoded into an image URL, so the browser just requests forty ordinary images and handles the loading itself. It is a 2023 build and I have not touched it since, which shows: while a batch renders, it reads you a haiku.