mispar

Hebrew gematria with all the classical counting methods — not just the standard count — plus Hebrew numerals in both directions. מספר, "number." Zero dependencies.

npm install mispar

Most gematria packages on npm do one thing: the standard count. But the tradition has many counting methods — final letters as 500–900, ordinal values, digit reduction, squares, spelled-out letter names, the atbash and albam ciphers — and they're exactly what gematria tooling needs.

import { gematria, letterValues, atbash } from "mispar";

gematria("חי");                    // 18
gematria("שלום");                  // 376
gematria("שָׁלוֹם");                 // 376 — niqqud stripped automatically

gematria("אלהים");                 // 86
gematria("אלהים", "gadol");        // 646 — final ם counts as 600
gematria("אמת", "katan");          // 9   — per-letter digit reduction
gematria("אמת", "katan-mispari");  // 9   — digital root of 441
gematria("א", "milui");            // 111 — אלף, the letter spelled out

atbash("בראשית");                  // the cipher as a string transform
letterValues("שלום");              // [{letter:"ש",value:300}, …] — build a calculator

Also exported: atbash(text) / albam(text) as string transforms, letterValues(text, method?) for per-letter breakdowns, METHODS, and MILUI_SPELLINGS (the default naming tradition). Pass variants through options.spellings, e.g. { ה: "הי", ו: "ויו" }. Letters you don't list keep the defaults.

import { hebrewNumeral, parseHebrewNumeral } from "mispar";

hebrewNumeral(18);                              // "י״ח"
hebrewNumeral(15);                              // "ט״ו" — never יה
hebrewNumeral(5);                               // "ה׳"
hebrewNumeral(5786);                            // "ה׳תשפ״ו"
hebrewNumeral(5786, { punctuation: "ascii" });  // `ה'תשפ"ו`
hebrewNumeral(786, { punctuation: "none" });    // "תשפו"

parseHebrewNumeral("ה׳תשפ״ו");                  // 5786
parseHebrewNumeral('תשפ"ו');                    // 786

hebrewNumeral takes an integer from 1 to 999,999: gershayim go before the last letter, a single letter takes a geresh, 15 and 16 are ט״ו and ט״ז, and thousands are written as their own letters followed by a geresh. An exact multiple of 1000 (5000 → ה׳) reads the same as a units numeral, so parseHebrewNumeral reads that back as 5. The parser accepts ׳ ״ or ASCII ' ", ignores niqqud, counts finals as their base letter, and doesn't enforce letter order (ער and רע are both 270).

(gematria("ט״ו") is also 15, but gematria ignores the thousands geresh, so use parseHebrewNumeral for years.)

• Niqqud, cantillation and other combining marks are stripped, as are invisible formatting characters (RLM/LRM, ZWJ/ZWNJ).
• Text is NFKD-normalized, so presentation forms (שׁ U+FB2A, wide letters, the אל ligature) count as plain letters; the Yiddish ligatures װ ױ ײ count as two letters.
• Geresh and gershayim (׳ ״, or ASCII ' ") are ignored inside a word, so רמב״ם is one word.
• Whitespace, maqaf (־), paseq (׀), sof pasuq (׃) and any other punctuation separate words. Only the positional methods (boneeh, haakhor) and the atbash/albam string transforms depend on where words break.
• Other non-Hebrew characters are ignored unless { strict: true }, which throws a RangeError on them.
• An unknown method throws RangeError. A non-string input throws TypeError.

By the same author: Gamatria — a gematria search engine over the Tanakh and Mishnah — and a fixed-income toolkit if your numerology runs to bonds instead.

Built by Moshe Malka — engineering leader in New York City. Studio work at Quentin.Code.

MIT © Moshe Malka