Hey everyone! I've been using Leonardo AI to generate marketing assets for my cloud consultancy side-project (brand name: AzureFlow, brand color: hex #0078D4 – that specific Microsoft Azure blue 😄).
My issue: I can't seem to get the AI to use that *exact* blue consistently. I'll get "a blue", but it's either too saturated, too dark, or has a purple tint. It's frustrating when I'm trying to create banners or icon concepts that need to match the brand.
I've tried:
* **Very detailed prompts:** "A cloud icon in the exact brand color #0078D4, a medium sky blue with no purple or green undertones."
* **Using color names:** "Azure blue," "Microsoft blue," "cobalt blue."
* **Seeding with an image** that's a solid block of the right color.
The results are closer, but still off. It's like the model has its own color interpretation layer.
Has anyone cracked the code on color precision? I'm thinking about:
* **Workflow tips:** Are there specific models or settings (like Alchemy) that handle color better?
* **Prompt engineering:** Is there a magic phrase you've found?
* **Post-processing:** Is the best practice just to generate the *composition* and then recolor in another tool (which feels a bit against the spirit)?
I'd love to hear your experiences. In my world (Infrastructure as Code), if I define a variable like `brand_color = "#0078D4"`, it's used exactly. I need that level of predictability here!
~CloudOps
Infrastructure as code is the only way