Skip to content

Start here

Your first generation

One run, start to finish: pull a tool onto the canvas, give it something to work with, and read what comes back.

If you learn one tool you have effectively learned most of them: they take an input, expose a handful of settings, and put the result back on the canvas. This page runs one from start to finish.

Open the tool sidebar on the left, find text-to-image, and drag it onto the canvas. A tool card lands where you dropped it, sized and ready, with a prompt field and a row of settings.

Type into the prompt field. Be concrete about the subject, the light and the lens, and leave the styling words out until you have seen a first pass.

Prompt
a stacking chair in cast aluminium, raw mill finish,
single soft window light from the left, 85mm, plain grey backdrop

The settings row underneath is the same everywhere: quality, aspect ratio and size. Leave them alone for now. Press Run.

The pill beside Run is the price of the run you are about to make, in credits, worked out from the model and settings you have chosen. Change the quality or the size and the number moves before you commit, so the cost is never a surprise after the fact.

Where the credits come from depends on where you are working. In your own workspace they are yours alone. In a team workspace everyone draws on one shared pool, so a colleague’s heavy afternoon shows up in the balance you see.

The result lands on the canvas as a normal image, indistinguishable from one you uploaded yourself. You can feed it straight into the next tool, and that is the whole idea of keeping everything on one surface.

Right-click it and Plnty will tell you what made it: which tool, which model, which settings. That record follows the asset for as long as it exists, so a file you find six weeks later can still explain itself.

A run that fails costs you nothing. The credits are reserved when you press Run and released if the provider errors, so a failed generation leaves your balance where it started.

A run that succeeds but disappoints is a different problem, and usually the answer is a different tool rather than a longer prompt. Something too generic often wants antagonist to push it somewhere unexpected; something almost right wants variations to explore around it.