What problem does it solve? Creating high-quality AI images with accurate text rendering, consistent characters, and reproducible results is difficult with plain prompts. This Skill connects Claude to Google's Gemini image models (Nano Banana series) through MCP, providing structured prompting workflows for marketing assets, UI mockups, and diagrams. ## Core Features & Use Cases - Image Generation & Editing: Create 4K images from text prompts, edit existing photos, and iteratively refine results using gemini_generate_image, gemini_edit_image, and continue_editing MCP tools. - Precision JSON Prompting: Translate plain-English briefs into structured JSON schemas (marketing_image, ui_builder, diagram_spec) that enable scoped edits—change lighting or camera angle without regenerating the entire scene. - Character Consistency & Text Rendering: Maintain the same character across multiple images using session history references, and render accurate in-image text for posters, logos, and infographics. - Use Case: A marketer needs a hero shot for a new beverage can. Describe the product in plain English, answer a few clarifying questions, receive a structured JSON spec, render it at 4K, then A/B test lighting variations by editing single JSON fields. ## Quick Start Ask Claude to generate a 4K image of your subject with a detailed description of style, lighting, and composition after configuring the nanobanana-mcp server with your Gemini API key.
