How It Works

The Greek Magical Papyri

The Greek Magical Papyri (Papyri Graecae Magicae, or PGM) is a collection of magical spells, hymns, and rituals written in Greek, Demotic Egyptian, and Coptic, dating from the 2nd century BCE to the 5th century CE. These texts were discovered in Egypt and represent a unique blend of Egyptian, Greek, Jewish, and Christian magical traditions.

The Bornless Ritual (PGM V. 96-172)

Also known as the Stele of Jeu the Hieroglyphist or the Headless Rite, this ritual is an invocation of the "Headless One" (Ἀκέφαλος, Akephalos), a powerful deity who created heaven and earth. The ritual combines Egyptian solar theology with Greek Neoplatonic philosophy and Jewish divine names, creating a syncretic magical practice typical of Greco-Roman Egypt.

The Vessel Inquiry (PGM IV. 154-285)

The Vessel Inquiry (λεκανομαντεία, lekanomanteia) is a divination ritual using a consecrated vessel of water as a scrying medium. A boy medium gazes into the water while the operator recites invocations to Helios and the divine light, summoning a presence to answer questions about the future.

Translation & Scholarship

The English translations are based on The Greek Magical Papyri in Translation, edited by Hans Dieter Betz (University of Chicago Press, 1986/1992), the standard scholarly edition. The phonetic reconstructions are based on modern scholarly consensus on ancient Greek pronunciation, with special attention to the voces magicae, the "words of power" that blend Greek, Egyptian, Hebrew, and invented barbarous names.

The Languages of Ancient Magic

The PGM texts are a linguistic tapestry, weaving together multiple ancient languages and magical traditions:

• Ancient Greek (Ἑλληνική): The primary language of the rituals, written in the Koine dialect spoken across the Hellenistic world. Greek provides the grammatical structure and narrative framework for most spells.
• Demotic Egyptian (Egyptian): The everyday script of ancient Egypt, used for native Egyptian spells and invocations to Egyptian deities like Ra, Thoth, and Anubis. Many Greek spells include Egyptian loan words transliterated into Greek letters.
• Coptic (Ⲙⲉⲧⲣⲉⲙ̀ⲛⲭⲏⲙⲓ): The final stage of the Egyptian language, written in Greek letters with additional characters. Coptic appears in later PGM texts and Christian-influenced spells.
• Hebrew/Aramaic (עברית): Jewish divine names and biblical phrases appear throughout the PGM, especially variations of the Tetragrammaton (YHWH) rendered as ΙΑΩ (Iao), ΣΑΒΑΩΘ (Sabaoth), and ΑΔΩΝΑΙ (Adonai).
• Voces Magicae (Barbarous Names): Invented or corrupted divine names that defy translation — ΑΒΛΑΝΑΘΑΝΑΛΒΑ, ΑΚΡΑΜΜΑΧΑΜΑΡΕΙ, ΣΕΜΕΣΙΛΑΜ. These "words of power" were believed to hold intrinsic magical force through their sound alone, regardless of meaning.

Text-to-Speech with Ancient Pronunciation

Every vox magica (magical word) is spoken aloud using ElevenLabs multilingual text-to-speech AI. When you hover over or click a word, the application:

1. Sends the phonetic spelling to the ElevenLabs API
2. Receives a streaming audio response in real-time
3. Plays the audio through your browser's Web Audio API
4. Analyzes the audio amplitude to drive the glow animation

Voice Settings

The voice is configured for dramatic, ritualistic delivery:

• Model: eleven_multilingual_v2 (supports Greek phonetics)
• Stability: 0.2 (low — allows dramatic variation)
• Style Exaggeration: 0.9 (high — theatrical delivery)
• Voice: Daniel (deep, gravelly male voice)

Phonetic Reconstruction

Ancient Greek pronunciation is reconstructed using modern scholarly consensus:

• Vowels: Pure vowel sounds (no diphthongization)
• Consonants: Aspirated stops (φ = "f", θ = "th", χ = "kh")
• Barbarous Names: Phonetic approximations based on Greek spelling

For example, ΙΑΩ (the Tetragrammaton in Greek letters) is pronounced "ee-ah-oh" rather than the modern "yah-oh" or Hebrew "yah-weh".

CSS-Only Papyrus Texture

The papyrus background is entirely procedural — no image files required. It's built using:

• 5 layered CSS gradients: Vertical and horizontal striations simulate woven papyrus fibers, plus three radial gradients create age mottling
• SVG feTurbulence filter: Adds organic noise and subtle warping
• feDisplacementMap: Breaks up the mechanical regularity of the gradients
• feColorMatrix: Desaturates slightly to simulate aged pigment

The result is a photorealistic aged papyrus texture that scales perfectly to any screen size and requires zero bandwidth for image assets.

Real-Time Audio-Driven Glow Animation

As each word is spoken, it glows in amber/gold tones that pulse in perfect synchrony with the audio. This is achieved through:

1. Web Audio API AnalyserNode: Extracts real-time amplitude data from the audio stream (60 times per second)
2. requestAnimationFrame loop: Updates CSS custom properties on each frame
3. GPU-composited CSS filters: drop-shadow() and brightness() create the glow without causing layout reflow

Glow Formula

The glow intensity is proportional to the audio amplitude:

• Blur radius: amplitude × 24px (0–24px range)
• Brightness: 1 + amplitude × 0.6 (1.0–1.6 range)
• Color: hsl(40, 90%, 55%) (amber/gold at 40° hue)

When audio ends, the glow fades smoothly back to the default state over 500ms using a linear decay animation.

Typography & Accessibility

The text uses a serif font appropriate for Greek/Coptic script, with a contrast ratio of at least 4.5:1 (WCAG AA compliant) against the papyrus background. Every interactive element is:

• Keyboard accessible: Tab to focus, Enter/Space to trigger
• Screen reader friendly: ARIA labels provide phonetic pronunciation
• Visually distinct: Focus indicators meet WCAG 2.1 AA requirements

Built with Kiro

This entire project was developed using Kiro, an AI-powered development environment that transforms how software is built. Kiro guided the project from initial concept through requirements, design, implementation, and comprehensive testing.

Spec-Driven Development

The .kiro/ directory in the source repository contains the complete development workflow:

• Requirements Document: 9 formal requirements with acceptance criteria
• Design Document: Architecture diagrams, data models, and 13 formal correctness properties
• Tasks Document: 22 top-level tasks with 56 sub-tasks, all completed
• Automated Hooks: Event-driven workflows that keep documentation in sync

Property-Based Testing

The application validates 13 formal correctness properties using property-based testing with fast-check:

1. Ritual sections always sort to canonical order
2. Phonetic mapping records are structurally complete
3. TTS request body always contains phonetic spelling
4. At most one concurrent TTS request per element
5. Glow state resets to defaults after playback
6. Glow values are proportional to amplitude
7. Usage log records contain all required fields
8. Session ID is always a valid UUID v4
9. Seed script execution is idempotent
10. Every VoceMagica element is keyboard-accessible
11. Keyboard interaction equals hover behavior
12. aria-label always equals phonetic spelling
13. Environment validation names every missing variable

Each property is validated with 100 random test cases, providing confidence that the application behaves correctly across all edge cases.

Test Coverage

The project includes 17 passing tests across 4 test files:

• Property-based tests: Validate universal correctness properties
• Integration tests: Verify Supabase, TTS API, and page rendering
• Unit tests: Test individual components and utilities

• Framework: Next.js 14 (App Router) with TypeScript
• Voice Synthesis: ElevenLabs multilingual TTS API
• Database: Supabase (PostgreSQL)
• Audio Analysis: Web Audio API (AnalyserNode)
• Animation: requestAnimationFrame + CSS custom properties
• Testing: Vitest + fast-check (property-based testing)
• Development: Kiro AI-powered development environment

This project is open source under the MIT License. The complete source code, including all Kiro development artifacts, is available on GitHub:

github.com/JustinMurray007/pgm-bornless-ritual

The repository includes setup instructions, architecture documentation, and the complete .kiro/ directory showing the spec-driven development workflow.