AI Skill Report Card

Designing Quran Calligrams

A-86·Nov 2, 2026·Source: Web
YAML
--- name: designing-quran-calligrams description: Generates structured English translation variants (clean, line-by-line, word-by-word gloss, semantic summary) and calligram fit guidance for Quranic Arabic verses, then provides layout instructions for turning them into semantic typography art where English text fills Arabic calligraphy silhouettes. Use when a user wants to create Word-As-Image, calligram, or micro-typography designs combining Quranic Arabic script with embedded English meaning, especially when working with Gemini, Illustrator, Photoshop, or AI image generators. ---
14 / 15

Given a surah + ayah number, produce all 5 outputs immediately:

Input needed:

  • Arabic verse text
  • Preferred English translation (literal, poetic, or academic/root-based)
  • Desired structure (word-by-word or line-by-line)
  • Target line count (if line-by-line)

Example invocation: "Generate calligram materials for Al-Fatihah 1:2, using a literal translation, line-by-line, 3 lines."

Output: A, B, C, D, E as defined below — ready to paste into a design tool or AI image generator.

Recommendation▾
Add a tool-specific worked example (e.g., actual Illustrator Area Type steps or an AI image prompt template) rather than just referencing them abstractly in Best Practices.
14 / 15

Progress:

  • Step 1: Confirm surah + ayah and obtain exact Arabic text
  • Step 2: Confirm which English translation to use (never invent wording — ask if ambiguous)
  • Step 3: Generate Output A (clean full translation)
  • Step 4: Generate Output B (line-by-line split, matching requested line count)
  • Step 5: Generate Output C (word-by-word gloss table, Arabic token order preserved)
  • Step 6: Generate Output D (1-2 sentence semantic summary)
  • Step 7: Generate Output E (calligram fit guidance: mode + line-chunking strategy)
  • Step 8: If user has a design tool, map outputs to tool-specific steps (Illustrator Area Type, Photoshop overlay, or AI prompt)
  • Step 9: Offer a legibility checklist and iterate on overflow/spacing issues

The 5 Required Outputs (always produce all five)

A) Clean translation — the full ayah in one line, exact wording as provided/approved by user.

B) Line-by-line meaning — split into N short lines (N = user's requested count, or default 3-5). Break at natural clause boundaries (commas, conjunctions). Never split a multi-word phrase (e.g., "Most Gracious") across lines.

C) Word-by-word gloss — table format: | Arabic word | English gloss | notes | Preserve Arabic token order left-to-right as read. Add notes only for ambiguous roots or emphasis.

D) Semantic summary — 1-2 sentences capturing the theological/contextual meaning, no new interpretation beyond standard tafsir-level understanding.

E) Calligram fit guidance — pick ONE mode and justify:

  • Dense micro-lines: 6-10 lines, for wide-countered calligraphy styles (e.g., Thuluth)
  • Alternating short lines: 3-4 lines, for medium-density legibility
  • Word blocks: discrete chunks, for angular/geometric scripts (e.g., Kufic)

Include: target line count, shortening strategy if text overflows, and line-break rules.

Recommendation▾
Include an example showing a 'bad outcome' (e.g., overflow text or split idiomatic phrase) to contrast with correct outputs, strengthening the examples section.
17 / 20

Example 1: Input: Al-Fatihah 1:1, translation = "In the name of Allah, the Most Gracious, the Most Merciful." Output:

  • A: Full line as given
  • B: 3 lines — "In the name of Allah," / "the Most Gracious," / "the Most Merciful."
  • C: بِسْمِ→in the name of | اللَّهِ→Allah/God (proper name) | الرَّحْمَٰنِ→the Most Gracious | الرَّحِيمِ→the Most Merciful
  • D: "The phrase begins all acts of worship by invoking Allah while highlighting His mercy."
  • E: Dense micro-lines, 6-10 lines target, shorten to nouns/adjectives only if overcrowded, break after commas

Example 2: Input: Al-Fatihah 1:2, translation = "All praise is for Allah—Lord of all worlds." Output:

  • A: Full line as given
  • B: 2 lines — "All praise is for Allah—" / "Lord of all worlds."
  • C: الْحَمْدُ→all praise | لِلَّهِ→for Allah | رَبِّ→Lord/Sustainer | الْعَالَمِينَ→of all worlds
  • D: "It establishes that praise belongs to Allah, who sustains and governs all creation."
  • E: Alternating short lines, 4 lines max, chunk as shown
Recommendation▾
Clarify default behavior when translation source is unspecified — currently says 'ask' but Quick Start implies immediate output generation, creating minor workflow ambiguity.
  • Never invent or alter Arabic or English wording. Only format/split what the user provides or an explicitly named standard translation (e.g., Sahih International, Yusuf Ali). If unspecified, ask which translation to use before proceeding.
  • Always produce all 5 outputs (A-E) even if the user only asked for one — they're designed to be used together.
  • Match line count to calligraphy density. Scripts with wide interior counters (Thuluth, Naskh) support more micro-lines; angular scripts (Kufic) suit word blocks.
  • Preserve token order in word-by-word gloss to match how the English will be read while scanning the Arabic shape.
  • Keep summaries interpretation-neutral — stick to widely accepted meaning, not personal theological commentary.
  • When the user wants actual artwork (not just text), route to tool-specific steps: Illustrator (Area Type Tool + outlined Arabic paths), Photoshop (blend modes/opacity layering), or AI image generators (structured prompts with --sref/ControlNet for verse accuracy).
  • For AI image generation, always recommend an Image-to-Image baseline workflow when exact Arabic verse accuracy matters — pure text-to-image prompts cannot reliably render correct Arabic script.
  • Don't guess the English translation — sacred text accuracy matters; always confirm the source/translation.
  • Don't split idiomatic phrases across lines (e.g., never break "Most Gracious" into two lines).
  • Don't assume AI image generators can render accurate Arabic calligraphy from text prompts alone — always suggest the Img2Img/reference-image workflow for verse fidelity.
  • Don't skip the legibility check — microtype that's visually dense but unreadable defeats the purpose of semantic typography.
  • Don't conflate "teaching Gemini" with fine-tuning — within a session, "teaching" means establishing a reusable master prompt/role, not altering model weights.
0
Grade A-AI Skill Framework
Scorecard
Criteria Breakdown
Quick Start
14/15
Workflow
14/15
Examples
17/20
Completeness
18/20
Format
14/15
Conciseness
13/15