Say what you want in plain words. Pick the model you are actually using, and get a prompt written the way that model expects, including the uncensored kind, which most prompt tools quietly refuse.
If you want more than this
Paid, mostly. Faster, no queue, and they keep your work.
Characters to talk to, with voice, calls and pictures in the chat
One trial generation before you make an account
The feed is open, the tool is behind a free account
Pony and Illustrious were trained on comma-separated booru tags and read a prompt as a list, quality words first. FLUX and Z-Image were trained on captions and read the same list as noise. Midjourney wants a sentence and then flags. Paste the wrong dialect into the wrong model and you get something that is technically what you asked for and looks nothing like it. That is the whole job here.
Most people do not know what is running under the tool they use, and that is fine. That option writes three things at once: a sentence version, a tag version and a negative prompt, with nothing model-specific in any of them. Paste whichever one your box asks for. It is not a fallback, it is a real setting, and it is the one that is picked when you arrive.
A still picture forgives a vague prompt. A clip does not: if you never say what moves, the model decides, and it usually decides on nothing. So for video we ask up to three short questions with tappable answers, and we skip any question you already answered in your own words. Every question also has a not sure, which takes a sensible default. You can clear the whole thing in three taps.
Some models will not run explicit work at all, no matter how the prompt is written. Rather than hand you something that bounces, we say so and write the strongest version that model will actually accept, and you can switch to one that will not refuse. No other prompt tool does this, which is strange, because it is the first thing you would want to know.
Questions
Yes, and per model. Where the model you picked is uncensored the prompt is written plainly. Where the model refuses that kind of thing we say so instead of pretending.
Pick I'm not sure, which is already picked. You get a sentence version, a tag version and a negative prompt, so one of them fits whatever you are pasting into.
Because a clip needs to know what moves, and most one-line ideas do not say. It asks at most three, it skips anything you already covered, and every question has a not sure.
Free, no account, no card. Same as everything else on this site.
More on what is listed here and why: how this list is built.