Using the ImageGen view allows you to work directly with DallE, the image generation AI we use to create images. This gives you greater control over the image outputs you receive.
How is ImageGen Different From ChatsterAI?
You can still create images in ChatsterAI simply by asking for an image.
The ImageGen view creates a space where only images can be generated, which is a powerful tool to define what you're looking for and get more defined image outputs.
When you generate an image with ChatsterAI, ChatGPT will read your prompt, and then rewrite it to send to DallE. Sometimes, things you asked for can be lost in translation (or things you didn't ask for are added). Using ImageGen allows you to go straight to DallE yourself, so you can be precise in your requests.
- ImageGen can only respond with images
- You do not need to say "make me an image of a 2d illustrated astronaut". You can just say "2d illustrated astronaut"
- ImageGen does not work as a conversation like ChatsterAI. For example:
- In ChatsterAI, you can say "make me an astronaut in space" and then follow up with, "make her floating in front of Jupiter."
- In ImageGen, each new prompt is separate from the others. So you would say, "an astronaut in space", then your next prompt would be "an astronaut in space in front of Jupiter".
Tips
- Use descriptive language. The more you give the AI to work with, the better your images will be. DallE can handle long and descriptive requests, though it may struggle with complex details.
- Get help from ChatsterAI. Struggling to describe what you want? Switch over to ChatsterAI, and let it know you'd like help writing a prompt for an image. Describe what you're looking for. Copy the output and enter it on the ImageGen screen to send it to DallE directly. This will give you more control. Of course, you could always just ask ChatsterAI to draw what it's written!
- No Custom Instructions. Unfortunately DallE is not able to parse custom instructions, even if they are tailored specifically to images.