AI Prompt Generator: Let Yamete Chat Write Your Image Prompts
Kryme
Yamete Chat is a creative assistant that works as an AI prompt generator: you describe a scene in plain words, it writes a prompt the model can read, renders the image, and keeps working on it with you in the same conversation.
What goes in, what comes out
You write the way you would tell a friend. The assistant turns it into tags. Here is an illustration of the kind of prompt it writes:
| You type | Kind of prompt it writes |
|---|---|
| A red-haired witch reading in a mushroom forest at night | 1girl, adult woman, red hair, witch hat, holding book, reading, forest, glowing mushrooms, night |
The exact tags change with the model and the request. It also adds negative terms, selects a style and settings, and renders the picture. From there, ask for a different outfit or a new background, or tell it to make the character move. The conversation carries on from the last result.
What does it do that a prompt box does not?
It knows the tag language
Most anime models were trained on Danbooru-style tags, and Yamete Chat writes in that vocabulary so you can skip learning it. If you want to learn anyway, our guide to facial tags and expressions is a good place to start.
It follows the rules of the model
Write "no cigarette" in a positive prompt and you may well get a cigarette, because the word is in the conditioning. A diffusion model draws whatever it reads. The prompt writer avoids negations on the positive side and moves unwanted things to the negative prompt.
It edits and animates
Send a picture or point at the last result, then describe the change, and the rest of the image is kept. Ask to animate a picture and the clip arrives in the same thread.
You can learn from the result
The prompt stays attached to each image, so you can read what was written and carry it over to Studio when you want to adjust every setting yourself.
How is it different from Roleplay?
Yamete Chat answers questions about generation, suggests ideas when you are stuck and turns your requests into renders. Roleplay is fiction with a character who stays in role and remembers your story. Both live on the same account.
Which language model does it use?
Yamete Chat runs on third-party language models selected for this work, with our own instructions and tooling on top. Your messages pass through those providers to generate replies, as the privacy answer in our FAQ explains.
Frequently asked questions
Can I see the prompt it wrote?
Yes. The prompt is shown with the image.
What does it cost?
Messages and generations use energy, and new accounts receive some to start. Yamete+ makes messages unlimited and adds monthly energy for images and videos. See the pricing page.
Can I ask for ideas instead of an image?
Yes. Ask for ideas and it will suggest some, and you can pick the one you want to render.
Getting better results
- Say who is in the picture first, then the outfit, the place and the mood.
- Make one change per message, since that is easier to steer.
- Read the prompt on the images you like and reuse what works.
- Use Studio when you want to control one setting exactly.
