We've improved image creation on WhatsApp with DALL-E 3. We're calling it Image Creation v2. In v1 you were able to create images once and would have to repeat that task over and over again, but you weren't able to refine it using prompts. But, now you can. Using ChatGPT-4 you can now refine an image using prompts. Here is how it works:
How to Create Images
Open your WhatsApp Chat with WizAI: Make sure you're in the right conversation with our WizAI bot.
Type the Image Command: To initiate an image creation, simply type the command !image followed by a detailed description of what you want to create. For example:
!image cute and realistic image of a puppy in a cafe in Paris
Replace the example prompt with your own description.
Send the Command: Press send, and the request will be processed by DALL-E 3, which will generate an image based on your description.
Receive Your Image: After a brief moment, your generated image will be sent to you in the chat.
Swipe to reply on the previous image that was created and type any refinement you'd like to do, e.g. "!image the dog should be a poodle". Don't forget to use !image before the reply.
Receive Your Image: After a brief moment, your updated image will be sent to you in the chat.
Tips for Writing Effective Prompts
Clarity is key: Clearly describe the subject, action, setting, and any specific details you want in the image.
Be descriptive: The more details you provide, the better DALL-E 3 can understand and create your image.
Use Adjectives: Descriptive words can help shape the final image. For example, instead of saying "cat," say "fluffy white cat with blue eyes."
Set the Scene: If you're looking for an image with context, describe the environment. For instance, "a futuristic cityscape at dusk" sets a clear scene.
Experiment: Don’t hesitate to try different phrasings or add more details if the first image isn’t quite what you expected.
Here are some prompts to help you get started:
!image a serene lake with a wooden pier at sunrise
!image a vintage car parked on a bustling city street at night
!image a whimsical illustration of a cat drinking tea in a garden
Tips and Best Practices
Originality: While DALL-E 3 can create images from existing concepts, try coming up with unique combinations or scenarios for truly one-of-a-kind creations.
Edit and Retry: If the first image isn't quite what you expected, tweak your prompt and try again. Sometimes small adjustments can make a big difference.
If the image doesn’t meet your expectations or if you’re having trouble with the command, consider the following:
Revise Your Prompt: Adjust the level of detail or try different descriptors.
Command Format: Ensure there are no typos in the command and that the format
!image [image prompt] is followed correctly.
Connectivity: Check your internet connection, as a stable connection is required for image generation.