Tips for Better Results

The robot converts text into an image based on training data provided by its creators. So, it may not fully meet your expectations, and you may need to try multiple input variations. You can try it on the labs.openai.com website or directly in the Bing browser. Also, remember that you will get different results each time for the same prompt.

Either way, be patient - creating an image takes about 20 seconds.

1. Specific Description:
Describe as specifically as possible what you want. Instead of just saying "tree," try to specify the type of tree (such as apple tree, pine, or bonsai) and the conditions it should have (for example, on a sunny day or in winter with snow).

2. Size, Proportions, and Lighting:
Specify the desired size and proportions of the object or scene. For example, "small red balloon" or "large city with clouds." You can also choose the type of lighting for the scene - soft light, dark shot, etc.

3. Colors, Textures, and Style:
Provide the desired colors and textures. You can use specific colors or characteristics, such as "matte red ceramic" or "glossy blue metal." Tell the robot in what painting style to create the image - in the style of Salvador Dalí, Vincent van Gogh, or photorealistically. Specify the technique as well - oil on canvas, watercolor, drawing, or photography.

4. Action or Context:
If you want to generate a scene with action or in a specific context, mention it. For example, "a group of people celebrating on the beach at sunset" or "a boy playing with a toy car."

5. What You Don’t Want in the Image:
If you have any specific requests or limitations, state them in the prompt. For example, "no human figures."

6. Experiment with Formulation:
Try different prompt formulations and observe how they affect the results. You can experiment with different phrasing, synonyms, and sentence structures.

7. Details:
Try different levels of image detail. Use prompts like "4k" or "8k" to specify the level of detail.