Independently, with no letter counting at all, the total word count of those same twenty-nine chapters is also 39,349. And at the level of the whole book: 114 surahs, 6,232 verses, 82,498 words, 332,519 letters, every one of them divisible by nineteen.
Reproduce it in three steps
git clone https://github.com/7430project/quran-fractal.gitpython3 verify.py. It needs nothing but Python itself.fractal_edition.txt to any capable AI, such as Claude, and ask it to check every claim.Get the files
FreePDFEnglish
The Fractal Quran, English Edition
An English rendering of the Quran assembled with AI assistance for one specific reason: to keep a single translator's theology, school, or era from steering the wording. The aim is the plainest defensible English for each verse, so the structure can be read without a thumb on the scale. Free to read and free to share.
FreeTXT1.4 MB
The Fractal Edition text and proof
The complete assembled source text followed by the full verification document: the counting procedure, the thirteen groups, and the result, all in one plain file. Upload it to any AI and ask it to check the claims, or read the proof directly. This is the exact file the verification script produces.
FreePythonSource
The verification code
The script and the source texts, in full. Clone it, run it with one command, and reproduce 39,349 on your own machine. No dependencies beyond Python itself. The point is to make breaking it easy, not impossible.
Everything here is free, and always will be. No payment is asked, and none is accepted.
File integrity
The proof file is published as is. Its SHA-256 lets you confirm the copy you downloaded is the copy the script produced.
fractal_edition.txt SHA-256
af4c00c0d4bfd843938b23cc999d965f9338b29f21a400fc6e0ed6e976b15f60
The text we count, and the frontier
Every count here runs on the standard Uthmani text, as fixed in the 1924 Cairo edition. The earliest manuscripts are older witnesses to that same text, and modern scholarship shows they descend from a single written archetype: see Dr. Marijn van Putten's Quranic Arabic (Brill, 2022) on the shared orthographic idiosyncrasies that prove it. The letter-level counting was made reproducible by the Quranic Arabic Corpus, developed under Professor Eric Atwell at the University of Leeds. Neither author endorses this project; their work is the record of what the text is, and the conclusions here are the project's own. It is worth saying why the text is worth counting at all: because it has stayed fixed. That preservation is the Quran's own claim about itself: We have sent down the Quran Ourself, and We Ourself will guard it (15:9).
The strongest version of the claim is to test the structure against that earliest documented form, letter for letter. That is the open frontier. The Manuscript Question lays it out, and the Team page is where to offer help.
Cite this work
If you reference the project in academic or written work, please cite a fixed version so a reader can reproduce exactly what you saw.
Caswell, M. (2026). The 74:30 Project: the fractal structure of the Quran's Muqatta'at (Version 1.0) [Data set and proof]. https://7430project.comCaswell, Michael. 2026. "The 74:30 Project: The Fractal Structure of the Quran's Muqatta'at." Version 1.0. https://7430project.com.@misc{caswell2026_7430,
author = {Caswell, Michael},
title = {The 74:30 Project: The Fractal Structure of the Quran's Muqatta'at},
year = {2026},
version = {1.0},
howpublished = {\url{https://7430project.com}}
}