The Manuscript Library What It Isn't Our Research Curiosities Media Collaborate Resources

c. 1404–1438 · 240 pages · Unknown language · Beinecke MS 408

The manuscript
no one has read.

600 years. Still waiting.

★ Start Here — What You Need to Know Experience it →
240 pages 0 translations 6 centuries of attempts 1 mystery
Explore
Latest:  Paper 7 v2.6 — Currier A/B CBMI discriminant (d=−1.01, p<0.001)  ·  New: Inside the Voynich, a mystery you can measure  ·  BPE analysis: 63 corpora, 35 language families  ·  Has it been solved?  ·  Researchers welcome

Sometime in the early 15th century, someone sat down and filled 240 pages of vellum with text and illustrations unlike anything seen before or since. The language is unknown. The script has no parallel. The plants don't exist. The stars don't match. Six centuries of scholars, cryptographers, and linguists have tried, and failed, to read it. Here you can read the full record, and the original, reproducible analysis we have added to it.

Original Research

What you won't find anywhere else

Beyond the history and the bibliography, this site reports first-hand computational analysis of the manuscript's text, measured on fixed, reproducible metrics and published openly.

ISR

An intermediate regime

Measured across 70+ writing systems, the Voynich lands in a real gap: far more organized than random gibberish, yet not quite matching a natural language.

d = −1.01

Two hands, by the numbers

The Currier A and B "dialects" separate at d = −1.01 (p < 0.001) on character-level structure: a measurable difference, not an impression.

100%

Reproducible by design

Every result is deterministic and re-runnable byte for byte. Code, data, and papers are published openly on Zenodo, not merely asserted.

See the findings → How we measure →
Where It Sits

Not gibberish. Not quite a language.

Measured across 70+ writing systems on fixed, reproducible metrics, the Voynich lands in a real gap.

Random gibberishNatural language

It is more organized than random noise, yet more rigidly patterned than any natural language we measured. That gap, an Intermediate Structural Regime, is the real puzzle: not a hidden message waiting to be unlocked, but a structure too organized to be noise and too rigid to be an ordinary language.

Experience the full story →

Inside the Manuscript

Six Sections. Six Mysteries.

The Voynich Manuscript is divided into distinct sections, each stranger than the last. All images are public domain — Beinecke Rare Book & Manuscript Library, Yale University.

Biological section - Bathing women

Biological

The Bathing Women

Nude female figures in interconnected pools and tubes. No theory has satisfactorily explained this section.

Astronomical section - Celestial wheels

Astronomical

The Celestial Wheels

Circular diagrams packed with stars, symbols and text. Some may depict constellations — but which ones?

Botanical section - Unknown plants

Botanical

The Unknown Plants

Over 100 plant illustrations — detailed, anatomically structured, yet matching no known species on Earth.

Cosmological section - The great map

Cosmological

The Great Map

A large fold-out page with concentric circles, landmasses, and figures. The most structurally complex page in the manuscript.

Zodiac section - Star women

Zodiac

The Star Women

Circular zodiac diagrams with female figures holding stars. The month names — written in a different hand — are the only readable text in the manuscript.

Pharmaceutical section - The vessels

Pharmaceutical

The Vessels

Rows of containers, jars, and tubes alongside plant parts and text. Possibly a formulary or recipe section — but of what?

Community Poll

What do you think the Voynich is?

Over 600 years of theories. Which do you find most convincing? Vote and see where the community stands.

Anonymous · No login required · Results update in real time

Thank you. Here's where the community currently stands:

Based on votes from this browser session.

Explore

Everything in One Place

From the manuscript's physical history to the most rigorous statistical analysis available — organized, sourced, and self-critical.

The Manuscript

History & Provenance

What it is, who owned it, how it was found, where it lives today. The physical object behind the mystery.

Explore →

Library

Everything Published

The most comprehensive curated bibliography on the Voynich — organized by theme, with critical notes on methodology.

Browse →

What It Isn't

Ruled-Out Hypotheses

Every eliminated hypothesis narrows the space. A growing map of what the evidence has ruled out.

See the list →

Curiosities

Surprising Facts

The NSA tried. So did WWII codebreakers. And 50+ people who claimed to have solved it. Verified, documented, strange.

Discover →

Media

What's Being Said

Curated coverage — rated rigorous, balanced, or sensationalist. Updated as new claims emerge.

Watch →

Collaborate

Join the Research

We are looking for corpus contributors, linguistic experts, and institutional partners. Full data access available.

Get in touch →
A computational perspective
📄 Paper 7 — Zenodo DOI 📄 Paper 8 — Zenodo DOI
"63 language corpora. 35 language families. Not one simultaneously matches the Voynich structural profile on all three metrics. The discriminant zone remains empty."

This site hosts original typological research by L. — a systematic comparison of the Voynich against known writing systems using BPE morpheme-length analysis. Among the findings: Tagalog is the closest natural-language match on one metric, but fails decisively on two others, and cannot maintain its elevated value across different texts.

The research is one perspective among many. It is reported with full transparency about what it can and cannot conclude.

Read the findings → See what we've ruled out → Download the data →
↑ Back to top