Prompt is still a bit hit or miss, and needs a reference image:
> A digitally sketched color doodle of a the person in the photo as an anime character in the style of the reference, no text, no embellishments. Eyes must be widely separated small black dots (no white). Keep the same hairstyle and clothing from the photo, a cute ditto-like character with gloopy hair (but not drippy!), long arms, full body shot, short legs. wide mouth
Just tested @GoogleAIStudio, asked it to build a @prompthero clone and the model literally prompt-injected itself trying to dodge its own safety filters?
✨ The consistency of Seedream 4 is simply mind-blowing. Switch accessories, faces, or styles, the results stay incredibly realistic.
Now live on PromptHero 🚀
The consistency of Seedream 4 is simply incredible. ✨ Just pick 3 clothes and a face, write a prompt, and this is the result. 👇
Now available on PromptHero, explore the next level of AI image generation. 🔥
Coral Protocol x @prompthero
We're working with PromptHero to build a MAS use case for their 700k registered users.
At Coral, we help businesses build scalable multi-agent systems 10x faster.
The agentic web is going mainstream.
Our partnership with PromptHero marks step one of that journey.
Notifs ON for more deets. 🔔