Type any Hebrew word — a name, a hope, a date — and search all five books for it hidden at equal skips, the way the codes have been read since Rabbeinu Bachya. Free, instant, and counted on the mesorah you choose.
יִשְׂרָאֵל — יֵשׁ שִׁשִּׁים רִבּוֹא אוֹתִיּוֹת לַתּוֹרָה"Yisrael" is an acronym: there are sixty myriad letters for the Torah — Megaleh Amukos, Ofan 186
The ink of a sefer Torah holds 304,805 letters by the received count — this page's codex a few spellings fuller, and either way barely half of six hundred thousand. The Pnei Yehoshua (Kiddushin 30a) asks the question outright, and the teaching stands anyway, because it counts a fuller Torah than the ink alone. Three classical mechanisms make up the difference:
The Alter Rebbe (Likkutei Torah, Behar 43d): count the אמות הקריאה — the vowel-letters the pronunciation carries but the ksiv omits. Every cholem without a vav, every chirik without a yud, is a letter of the word's full form. Written out, the count climbs toward 600,000 — and the mikra, the Torah as read, is itself Sinaitic (Nedarim 37b).
The Torah pre-exists as black fire on white fire (Yerushalmi Shekalim 6:1). The parchment-forms between the ink are letters not yet legible, to be revealed l'asid lavo — and the halacha already guards them: a letter whose surrounding white is broken invalidates the scroll (מוקף גויל, Menachos 29a). Sixty myriad counts black and white together.
Every letter decomposes into elemental strokes — an aleph is two yudin and a vav, and Rabbi Akiva expounded mounds of halachos on each thorn (Menachos 29b). Counted by their parts, the letters of the Torah approach 600,000. And a yud is a Yid.
Sixty myriad letters against sixty myriad root-souls of Yisrael (Tanya ch. 37): every Jew has a letter, and a scroll missing one letter is pasul entirely — the engine behind the Rebbe's Os B'Sefer Torah campaign, and the reason this project carries the name. Note the acronym's lamed: לַתּוֹרָה — sixty myriad letters for the Torah, not merely in it. What the ink shows is the beginning of the count, not the end of it.
The Torah is one. What differs, at a handful of letters, is the mesorah of its writing — the sofrim’s transmitted spelling traditions. Every figure on this page — the census, the middle, the frequencies, the grids — is recomputed from the letters of the mesorah you choose. Pick a tradition, or flip any single site to build your own. Both readings at every site are kosher texts by the standards of the poskim (Mishnas Avraham 32:3; Kaf HaChaim 143:34; Yechaveh Daat 6:56).
Fourteen switches: nine community letter-sites, three readings attested for the Aleppo Codex alone, and two word-divisions that move the word count without touching a letter. Canonical totals assume the received 304,805-letter base; live totals here run on the Leningrad base until the received correction layer ships. Twelve binary letter-sites give 4,096 possible letter-sequences — a find that survives all of them survives every tradition.
Type a Hebrew word and the engine scans the Torah, as counted by the active mesorah, for equidistant letter sequences, forward and backward, across the skip range you set. Final letters match their base forms, as the codes literature counts them. Any find whose span crosses a contested site carries the tradition-sensitive mark — switch shitos above and run it again to see whether it stands.
The early sages were called soferim — counters — because they counted every letter of the Torah (Kiddushin 30a). The figures below are not typed into this page; the browser derives them from the active shita's stream.
| Sefer | Letters | Words | Verses |
|---|
Words are graphic words as inked — maqqef-joined words counted separately. Verse totals follow this codex's versification, which divides the Aseres HaDibros differently than printed Chumashim; the letter-stream is unaffected. The Gemara itself records that verse division varied — in Eretz Yisrael, Shemos 19:9 was read as three verses (Kiddushin 30a).
לְפִיכָךְ נִקְרְאוּ רִאשׁוֹנִים סוֹפְרִים — שֶׁהָיוּ סוֹפְרִים כׇּל הָאוֹתִיּוֹת שֶׁבַּתּוֹרָה. שֶׁהָיוּ אוֹמְרִים: וָא״ו דְּגָחוֹן — חֶצְיָין שֶׁל אוֹתִיּוֹת שֶׁל סֵפֶר תּוֹרָה. דָּרֹשׁ דָּרַשׁ — חֶצְיָין שֶׁל תֵּיבוֹת. וְהִתְגַּלָּח — שֶׁל פְּסוּקִים.Kiddushin 30a
Measured on the active text, the midpoint falls letters before the vav of גחון (Vayikra 11:42) — the letter the soferim named as the Torah's belly-center, written large in every scroll (Sofrim 9:2). The Gemara saw the question coming. Asked to settle which half the vav belongs to, Rav Yosef answered for all generations:
אִינְהוּ בְּקִיאִי בַּחֲסֵירוֹת וִיתֵירוֹת — אֲנַן לָא בְּקִיאִינַן."They were expert in the defective and plene spellings; we are not expert." — Kiddushin 30a
Letter-codes did not begin with computers. The entries below are counted readings cited in the classical literature — Chazal, the Targum, Rabbeinu Bachya, the Baal HaTurim — each one rendered live from this page's text, with the counting system named. If a claim did not verify letter-for-letter against the text, it is not on this page.
R' Chaim Michoel Dov Weissmandl drew the Torah as grids by hand and found that from the first tav of בראשית, every 50th letter spells תורה — and again at the opening of Shemos. In Vayikra, from the first yud, every 8th letter spells the Shem Havayah. The grids below are built from the text of the active mesorah, one row per skip, so the sequence stands as a column.
Published posthumously by his talmidim in Toras Chemed (1958). The first documented letter-skip reading is far older — Rabbeinu Bachya (d. 1340) presents a four-letter sequence at 42-letter intervals at the opening of Bereishis.
In the 1980s Prof. Eliyahu Rips of the Hebrew University took up Weissmandl’s grids with computers, and Aish HaTorah’s Discovery Seminar carried the findings to hundreds of thousands worldwide. The cards below cite that canon — and, per this page’s rule, every figure is recomputed live from the active mesorah before your eyes. The matrices follow the arrangement Rips fixed and Aish made famous \u2014 one row per skip, the code standing as a single column \u2014 available on any grid via the Aish wrap button. Where the popular retelling and the text disagree, the text wins, and we say so.
For the technical case, Prof. Robert Haralick\u2019s torah-code.org \u2014 the site Jewish Action itself directs readers to \u2014 and the book he wrote with Prof. Rips z\u201dl and R\u2019 Matityahu Glazerson, Torah Codes: A Glimpse into the Infinite (2005). For the primary experiments, Doron Witztum\u2019s research site, and the papers themselves: Witztum\u2013Rips\u2013Rosenberg, Statistical Science 9:3 (1994) 429\u2013438, with the critique at 14:2 (1999) 150\u2013173 \u2014 read both. For the hashkafah and the poskim, Jewish Action\u2019s Torah Codes and Kiruv Rechokim and Torah Codes Revisited. For the method\u2019s roots, Moshe Katz\u2019s CompuTorah.
And a standing invitation: R\u2019 Glazerson\u2019s seforim alone hold hundreds of published tables. Bring any of them here \u2014 type the main term, set its skip, choose the mesorah \u2014 and this page will rebuild the matrix in front of you, or tell you honestly that it doesn\u2019t reproduce. That is what this bench is for.
Beyond the contested sites in the selector above, roughly a hundred plene/defective spelling differences separate the Leningrad Codex from the received text, every one documented in the masoretic literature, none changing a word's meaning. The Rishonim built the received text exactly this way: the Ramah gathered the oldest accurate codices and followed the majority, as the sages did with the three scrolls found in the Azarah (Yerushalmi Ta'anis 4:2). The attestation record per site:
Any letter-skip whose span crosses a switched site shifts by one letter between traditions. Sixty Myriad's search engine flags such finds as tradition-sensitive — a first for the genre.
Gematria is a sanctioned instrument — middah 29 of the Baraisa of Thirty-Two Middos, called "appetizers of wisdom" in Avos 3:18. The Ramak catalogs these systems in Pardes Rimonim, Gate 30. Type any Hebrew word: