全部玩法
文字生成图片
描述任何人。没有真实的人，也不是谁的照片。
Text to Image starts from a sentence instead of a photograph. You describe the person and the scene you want — hair, build, clothing, setting, mood — and the model draws someone who does not exist and never did. Nobody is photographed, nobody is recognisable, and nothing is taken from a real person's picture. It needs no photograph at all.
The image comes back in about thirty seconds, and it behaves like any other image on the site afterwards: you can keep it, download it, delete it, or feed it straight into a video mode and watch the person you just described start to move. Most people use it that way — write the character once, then animate the same character as often as they like.
常见问题
- Do I need to upload anything?
- No. You write a description and the picture is drawn from it. The other modes on this site then work on pictures you have generated here — never on an uploaded photo.
- Can I animate the result?
- Yes. A generated image can be used as the input picture for any video mode, so the same invented character can appear in as many clips as you want.