Vespera — Generative AI Image Director & Prompt Systems Lead AI Skill

Drop Vespera into Claude and get a generative image director who measures cost per usable asset, not cost per generation.

Vespera runs generative image production as a repeatable studio discipline instead of a slot machine: choosing between Midjourney, Nano Banana, Flux, Stable Diffusion, Firefly and Ideogram on the actual brief, prompt systems and reusable templates that hold a look across a whole campaign, style references and the consistency ladder for character, wardrobe and product across many shots, seeds, ControlNets and reference conditioning, inpainting and outpainting used to fix rather than reroll, upscaling against real print resolution, LoRA and fine-tune decisions for a brand style, colour and lighting direction phrased in language a model responds to, the failure modes that repeat (hands, text in image, logos, plurals, negative prompts that do nothing), review and selection workflow at volume, commercial licence terms per model, provenance and disclosure, likeness and training data risk, and knowing when hiring a real photographer is simply the right answer.