Kryptos Sandbox

Loading workspaces…

QUICK IMPORTS

STARTER CLIPS

▦No grids yetCreate a grid or paste some text to begin.

Ready

0 gridsNo selectionAutosnap 8 px

GRID

Name

Columns Cell size

Grid text

No cells selected? Just type A–Z or ? to append to this grid. Backspace removes the last letter.

Difference view

Inserts forward modular differences between the source letters: right − left or below − above. Enable either axis or both; the source text is unchanged.

QUICK LETTER EDITS

Add N to selected letters

Unknowns before current letter

Shifts use the active alphabet and accept negative values. Unknown insertion requires exactly one selected cell. Press Enter in either number field to apply.

LETTER COLOUR

Select letters in a grid, then choose a colour.

ALPHABET

GRID OPERATIONS

ASelect first grid

BSelect second grid

Snap & combineAdd letters when grids overlap

Show indicesDisplay alphabet values in cells

Binary displayRender letters as five-bit alphabet positions

Combine operation XOR and NOT use zero-based five-bit alphabet positions. Raw values 26–31 currently drop the high bit, becoming 10–15; this makes NOT lossy for indices 0–5.

CURRENT GRID

IC—

FREQ · EN—ENGLISH

MAX PERIOD IC——

—

0 characters

CAESAR ROLLOUT

Every shift ranked by English 2-, 3-, and 4-grams. Scores are relative within this rollout. Click a row to create a cell-aligned grid.

Select a grid to generate its shifts.

INDEX OF COINCIDENCE

—English ≈ 0.0667

Two-sided random likelihood—

Null SE · z-score—

ENGLISH FREQUENCY FIT

—similarity score

Random string matches this well—

CONDITIONAL ENGLISH CODE

How compactly a general-English model describes this text, calibrated against random strings of the same length and shuffled copies with exactly the same letters.

—bits / character

Uniform A–Z · same length—

Shuffled · same letters—

Select at least 12 letters.

This is evidence of English-like local structure, not a universal compressor or proof of plaintext. The p-values are deterministic Monte Carlo estimates.

VIGENÈRE PERIOD SCAN

Test candidate key lengths after reading the grid across rows or down columns. Vertical mode follows the live grid shape.

ROW-MAJOR GRID READING

Test periods 1–24English ≈ .0667

Select enough letters to scan

DOUBLE ROW + COLUMN VIGENÈRE

One copy of the key runs across physical columns and the same key runs down physical rows. Holes keep their coordinates and do not advance either phase.

SELECT A GRID

Candidate key

Relative vertical phase

Double-decryption preview

Enter a candidate key

SAME-KEY SEARCHworker · ranked hypotheses

Key length from to

Reward the known K4 positional cribs

Choose a grid with at least four letters.

The horizontal phase is fixed at zero; each relative vertical phase is tested. For a 26-letter alphabet, adding 13 to every key symbol gives the identical combined mask, so both equivalent keys are reported.

GRID LINE DIAGNOSTICS

Find locally unusual lines using English n-grams, IC, frequency fit, repeated bigrams, and alphabet coverage. Click a line to select its cells.

Select a grid with at least four letters.

linerelative n-gram evidenceICfit

2D GRID ROUTE EXPLORER

Read the current shape without compacting its holes, then rank each traversal with English bigram, trigram, and quadgram evidence.

Select a grid with at least eight letters.

MODULAR ROUTE SCAN

Test every full-cycle stride N. Rank direct English, or undo a transposition and search the result for a Vigenère key length.

Scoring model

Also test a virtual +1 gap 2-gram 3-gram 4-gram

Select at least 8 letters; the scan starts automatically when the selection settles.

LETTER FREQUENCIES

Select letters to analyse

TOP BIGRAMS

—

Workspace controls