Stable Diffusion Image Bot

You need a picture for a post, an avatar or just out of curiosity, but you cannot draw. In this bot you describe the idea in words and the Stable Diffusion neural network draws it for you 🎨

The bot calls itself a free image creator that generates and edits images and writes prompts. It is an unofficial bot, with no connection to the developers of the model itself.

Four ways to start

The welcome message explains right away what you can send. Choose whatever fits your task:

1. A text prompt: describe what you want to see and get a picture from scratch.
2. One image: the bot processes your picture.
3. An image with a caption: a photo plus text saying what exactly to change.
4. Three to five photos as a group: for tasks where several shots must be uploaded at once.

Knobs for the curious

There is more control here than in simple generators. The /steps, /scale, /strength and /negative commands change generation parameters, and /reset puts everything back.

If you have no idea what these mean, that is fine: you can leave them alone. Still, it helps to know what hides behind the words 🔧

• Steps: how many passes the network takes to refine the picture. More steps usually give more detail but take longer.
• Scale: how strictly the model follows your text. High values match the prompt closer, though the image can look harsher.
• Strength: when editing a finished image, shows how much it will change.
• Negative: a list of what must not appear, such as extra fingers or blur.

The queue and its status

Generation takes computing power, so requests wait in a queue. The /status command shows its state, which tells you whether to wait for the result or come back later.

If a picture does not turn out, do not worry. Reword the prompt and try again 🖼

Writing a prompt that works

A good prompt answers four questions: what is in the frame, where, in what style and under what light. "An old lighthouse on a cliff, sunset, oil painting" works better than plain "lighthouse".

Add details step by step. Check the main idea first, then refine the style and small things.

If the network keeps drawing something unwanted, put it into the negative prompt with /negative. That is easier for removing blur, stray text or distorted hands than rewriting the whole prompt.

Rights and risks

Images made by a neural network may resemble existing works or real people. For commercial use, check the image usage rules in advance.

Do not upload someone else's photos without the person's consent, and do not use the bot for forgeries or deception. Avoid sending personal documents or pictures you would not want exposed 🔒

Support and donations

The bot has a /donate command, which is voluntary support for the project. Before any payment, check where the money goes and what it is for, and do not send more than you are ready to spend.

If something works wrong, try the /help command, which describes the features. The profile lists a support contact for service questions.