Trying a Prompt After You Buy

Run it from your library

Open any purchased prompt in your Library and hit Use this prompt. Fill in the prompt’s variables with your own details, and the page shows the exact credit cost before you run — some models also offer extra options, like a priced high-resolution tier for images or a cheaper 5-second preview cut for video. Hit Run and the result generates on the real model, appearing right on the page.

Everything you generate is kept in an outputs gallery on that same page, with the exact prompt text each run used — and downloads of the full-size files.

A few prompts can’t be run in-platform — models we haven’t wired up yet, and certain configurations (like image-to-image, or video with audio). Copying the prompt to use in your own tools is always available regardless.

How long does a run take?

Runs aren’t instant. When you hit Run, your result is generated fresh by the AI model, so the wait is the model’s own generation time — it comes from the model provider, not from a queue on our side. Many image models finish in seconds and most within a minute or so; Midjourney usually takes a couple of minutes. Video is the long one — expect several minutes for a video run. Any model can be slower when its provider is busy.

The page tracks progress for you, so there’s nothing to refresh — and no need to babysit it either. You can leave at any point; the run keeps going, and the result will be waiting in the prompt’s outputs gallery when you come back. If a run fails or never finishes, its credits are refunded automatically.

Credits — how you earn and spend them

Every paid prompt includes bonus credits, shown right on its page next to the price — higher-priced prompts include more, and every paid prompt includes at least 100. Bundles and seller subscriptions grant credits too. Credits are a perk of paying for a prompt, so a prompt you receive free through a promotion doesn't include them. Credits from all your purchases pool into one balance, and your Credits page shows your balance and every entry in your history.

Which models are currently supported?

The table below is the live list — every model shown there can run in-platform, and it updates automatically as we add support for more. If a prompt’s model or version isn’t runnable yet, its page in your library says so, and copying the prompt to use in your own tools always works.

Why image and video models? A visual prompt produces a complete result in a single run — perfect for spending a few credits to see it deliver. Text prompts work differently: they’re built for conversation, where you iterate with the model over a session and steer it as you go, and a single fixed reply would short-change what you bought — so text prompts are best used by copying them into your own chat with the model they’re written for.

Video runs are priced per 5 seconds of output, so a longer prompt costs proportionally more — and prompts longer than 5 seconds also offer a cheaper preview option that renders a 5-second cut of the same prompt, so you can watch it deliver before spending on the full length.

What runs cost

Prices are live — this is exactly what a run costs right now, per model.

Run pricing will appear here once in-platform runs are live.