Grandpré cipher

The Grandpré cipher is a classical homophonic substitution that numbers every letter by its coordinates in a grid of keyword rows. You supply the words that fill each row of the checkerboard, paste plain text or a digit-pair stream, and choose encrypt or decrypt. Encryption walks the Latin letters of the message and replaces each one with the two-digit row-column label of the first matching cell in row-major order; decryption collects digits, reads them as pairs, and looks each pair back up on the same grid. Because the same letter can appear more than once among the keyword rows, the system is homophonic in principle—repeated plaintext letters need not always share one code—while this implementation stays fully deterministic by always taking the first cell. When the keyword list does not cover the full alphabet, the response still succeeds for letters that are present and reports incomplete grid status with the list of missing letters so you can enlarge the square.

How Grandpré keyword rows become coordinate pairs

A Grandpré square is built from words laid out one per row so that every cell holds a single Latin letter. Classical descriptions prefer a ten-by-ten board whose ten words also make the first column spell an eleventh word, but any rectangular set of keyword rows with equal letter counts works the same way: strip non-letters, uppercase the rest, and demand that every row ends with the same width between two and ten columns, with at most ten rows.

Axes are labeled from one upward; when a side has ten cells, the tenth label is the digit zero, matching the textbook 1–9–0 convention. Encryption replaces each plaintext letter with the two-digit code formed by its row label followed by its column label. If the letter appears several times in the grid, this tool always selects the first match scanning left to right and top to bottom, so identical inputs always produce identical digit streams without randomness.

Decryption reverses the table lookup: digits are harvested from the input in order, then consumed two at a time as a row label and a column label. An odd digit count, a label outside the active axis set, or a pair that does not address any cell fails immediately. Non-letters in the plaintext are skipped on encrypt so the ciphertext stays pure coordinate digits; non-digit characters on decrypt are ignored while the digit stream is assembled.

There is no modular alphabet shift and no secondary additive key—only the open keyword grid—so the cipher stands or falls on how completely and carefully those rows are chosen.

Where Grandpré tools show up in real work

Introductory cryptography and puzzle-club sessions still assign Grandpré-style grids so students practice homophonic coordinate substitution before they meet machine ciphers; a single trusted oracle that encrypts and decrypts from explicit keyword rows speeds demos and auto-graders. Capture-the-flag warm-ups and geocaching challenges hide digit-pair messages beside a published word square, and solvers want a fast, predictable first-match encoder rather than a general cryptanalysis suite.

Content and support teams sometimes lightly encode spoilers, quiz answers, or internal labels as coordinate pairs over a shared word list and need the reverse path without installing a language-specific crypto library. If you need a fixed five-by-five Polybius square without keywords, a Playfair digraph square, a Nihilist additive layer, Caesar or Vigenère letter shifts, or automatic recovery of an unknown Grandpré board from ciphertext alone, treat those as separate search intents and landings rather than overloading this keyword-row coordinate cipher page.