amw-excalidraw-illustrations

Generate hand-drawn Excalidraw-style concept illustrations as PNGs via the Gemini image API.

Updated Jul 8, 2026
One-click install
npx skills add https://github.com/Emasoft/ai-maestro-webdesign --skill amw-excalidraw-illustrations-emasoft
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: amw-excalidraw-illustrations
Source: https://github.com/Emasoft/ai-maestro-webdesign/tree/main/skills/amw-excalidraw-illustrations
Command: npx skills add https://github.com/Emasoft/ai-maestro-webdesign --skill amw-excalidraw-illustrations-emasoft

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires Pillow, and includes scripts (resource) and references (resource) components.

What problem does it solve? Creating consistent hand-drawn whiteboard-style concept illustrations for slides and educational material normally requires a human illustrator; this Skill produces them through the Gemini API with a fixed visual style, correct in-image text, and a mandatory verification loop. ## Core Features & Use Cases - Style-anchored generation: Embeds two shipped reference PNGs in every Gemini call so output stays within a consistent Excalidraw rough-sketch aesthetic across a whole deck. - Text-correctness techniques: A seven-section prompt template with letter-by-letter spelling blocks and framed-text rules keeps labels legible and correctly spelled; a two-phase Pillow overlay fallback fixes stubborn misspellings locally. - Gated cost control: Requires GEMINI_API_KEY and explicit per-call user consent, with aspect-ratio selection (16:9, 1:1, 4:3) asked before every paid call. - Use Case: A teacher needs a two-panel illustration comparing Realism and Naturalism for a slide deck; the Skill asks for the aspect ratio, builds the structured prompt with spelled-out author names, calls gemini-3-pro-image-preview, verifies every word in the PNG, and saves a descriptively named file. ## Quick Start Ask the assistant to create an Excalidraw-style hand-drawn illustration of your concept, for example: "Make an Excalidraw-style illustration of the OSI network stack in 16:9 for my slides."

Frequently Asked Questions about amw-excalidraw-illustrations

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I generate an Excalidraw-style illustration with Gemini?▼

Set the GEMINI_API_KEY environment variable, then invoke the skill with a concept and aspect ratio. It builds a seven-section prompt with two reference images as style anchors and calls gemini-3-pro-image-preview, returning a verified PNG.

How do I fix misspelled text in AI-generated illustrations?▼

Include a letter-by-letter spelling block in the prompt, which cuts text errors roughly tenfold. If two attempts still fail, use the two-phase fallback: generate a visual-only image with the flash model, then overlay text locally with Pillow and the Caveat font.

What is the difference between gemini-3-pro-image-preview and gemini-2.5-flash-image?▼

The pro model is the default because it renders in-image text accurately enough for educational material. The flash model is cheaper and faster but produces unreliable text, so it is reserved for visual-only composition iteration in the two-phase fallback.

Does this skill require an API key or paid access?▼

Yes, it requires a GEMINI_API_KEY from Google AI Studio, and every Pro call is billed to the user's Gemini quota. The skill aborts if the key is missing and asks for explicit consent before each paid call.

When should I not use Excalidraw-style AI illustrations?▼

Do not use it for photorealistic images, vector-flat art, characters, mascots, logos, or technical diagrams like flowcharts and architecture graphs. Those requests are refused and routed to the appropriate diagram or SVG skills instead.