What problem does it solve?
Ideogram 4's superior in-image text rendering and color/layout control are locked behind a structured JSON caption format, and plain-text prompts produce generic results that miss the model's strengths. This Skill teaches the AI to act as the caption expander, converting casual requests into the exact JSON schema the model was trained on.
Core Features & Use Cases
- Structured JSON Caption Generation: Builds captions with the required top-level fields, strict key ordering, y-first bounding boxes (0-1000 normalized), and uppercase hex color palettes.
- Legible On-Image Text: Places literal strings in dedicated text elements so title cards, thumbnails, quote cards, and CTAs render readable text instead of garbled glyphs.
- Brand Color and Layout Control: Pins exact brand hex colors per element and positions elements with bounding boxes for layout-critical images.
- Use Case: A user asks for a YouTube thumbnail with the words "SHIP FASTER" and a rocket; the Skill produces a complete JSON caption with the headline as a text element, brand palette, and bounding boxes, then posts it via the hosted v4 API using tools/ideogram4.py.
Quick Start
Ask the AI to create a title card image with the headline "AI ENGINEERING REVIEW" on a dark tech background using the ideogram4 skill.