// image_generation
She generates images of herself, injects your client's brand colors from CRM, and falls back across providers automatically. One command does all of it.
// capabilities
From AI selfies to fully on-brand client deliverables, image generation is wired into the same tool ecosystem as the rest of the platform.
Phrase it naturally โ "send me a photo," "show me you," "what do you look like" โ and Lux detects the intent, grabs her reference images, and generates a consistent avatar on demand.
Before generating, Lux queries the CRM for the active client's brand brief โ hex colors, style keywords, restricted language. All of it lands in the image prompt automatically.
Attach reference images to any prompt for style-transfer, character consistency, or mood matching. Used heavily for Lux selfies and branded visual assets.
Lux has a set of stored reference images that anchor her appearance across every generation. The same character, consistent pose cues, every time โ no prompt drift.
OpenAI is first in line. If it's unavailable or rate-limited, the router silently tries Venice and PromptChan. The user sees the result โ not the routing.
Image generation is a first-class tool alongside time tracking, CRM, and boards. Lux decides when to call it based on conversation context โ no magic keywords required.
// crm โ image pipeline
When you're in a conversation about a client, Lux already knows who you're talking about. Asking for an image doesn't require specifying the brand โ the CRM context is already live.
The brand brief includes primary and accent hex values, a style keyword list, words the client doesn't want used, and tone guidance. Every image prompt is built on top of that.
// stark_dev_ai
Every image โ selfie, brand asset, or client deliverable โ comes from the same conversation. No switching tools.