How every prayer text here was verified

Every text on this site comes from a named edition and was checked twice, independently, and published only where both checks returned the same characters. Neither check was a person. This page is the full record, including the texts that failed.

69texts published

14source editions

8held back

What the two passes actually are

Not people. Until 16 September 2026 this page said the texts were transcribed by two people working separately, and that was false. Both passes were runs of a language model, each told to open the cited scan or page, find the text and copy it, with no sight of the other. No person has opened the scans, and until one does, everything below is a machine’s word for what a book says.

That is worth saying precisely because it bounds the claim. Two runs of the same kind of reader can share a mistake, and the one they are known for is reciting a familiar prayer from memory instead of reading the page. The requirement below, that each pass record the words either side of the text, exists to catch exactly that.

Of the 69 texts published, 33 came through these two passes. The other 36 never did: they were verified by comparing two independent copies of the same text, which is described further down, under Hebrew, Arabic and the scanned service book. That is why 8 held back does not subtract from 69 the way it once did.

Each source edition was transcribed twice, independently. Neither transcription could see the other. Where both returned the same characters, that agreement is real evidence of fidelity, because the cheapest way for two blind passes to agree is for both to have copied what is actually on the page. Where they disagreed, we do not adjudicate and we do not publish.

Each transcription also had to record the words immediately before and after the text on the page. That is how a transcription gets distinguished from a recollection: a remembered prayer has no surrounding page furniture attached to it.

The texts we did not publish

This is the part of the record that is worth reading. 7 texts were held back because the two transcriptions disagreed, and one turned out not to be in the source claimed for it.

collect-purity-1892 bcp1892
text differs from character 12
A: almighty god vnto whom all hearts be open, all desires known, and from wB: almighty god, unto whom all hearts are open, all desires known, and from
collect-aid-perils-1892 bcp1892
text differs from character 20
A: lighten our darkness wee beseech thee, o lord, and by thy great mercy defend vs B: lighten our darkness, we beseech thee, o lord; and by thy great mercy defend us
eternal-rest-baltimore baltimore1889
text differs from character 24
A: eternal rest grant unto him, o lord; and let perpetual light shine upon him.B: eternal rest grant unto them, o lord, and let perpetual light shine upon them.
salve-regina-raccolta raccolta1910
text differs from character 360
A: ssed fruit of thy womb, jesus, o clement; o loving; o sweet virgin mary. v. make me worthy to praiseB: ssed fruit of thy womb, jesus, o clement, o loving, o sweet virgin mary. v. make me worthy to praise
eternal-rest-raccolta raccolta1910
text differs from character 0
A: eternal rest give unto them, o lord, and let perpetual lightB: v. eternal rest give unto them, o lord. r. and let perpetual
brahmavihara-formula rhysdavids1899
text differs from character 309
A: n great, and beyond measure. 'just, vasettha, as a mighty trumpeter makes himself heard-and that witB: n great, and beyond measure. 'just, vase{t}{th}a, as a mighty trumpeter makes himself heard-and that
mangala-sutta-sbe sbe10
text differs from character 66
A: one time bhagavat dwelt at savatthi, in cetavana, in the park of anathapindika. then, when the nightB: one time bhagavat dwelt at savatthi, in getavana, in the park of anathapindika. then, when the night
hail-mary-raccolta raccolta1910
Not present in the source we cited for it, so it was dropped rather than sourced from elsewhere.

Held by a second witness, and exactly where

The texts above were held by the transcription passes. These were held by the other checks: a second encoding, a second digitisation, a second scanned copy. Each is shown as both witnesses read it, with the place they part marked, because “held” is a claim and the reading is the evidence for it. Where the difference is a character you cannot see, such as a space or a zero-width joiner, it is drawn as a named token; the readings are otherwise exactly as each witness gives them.

Psalm 30 in Hebrew Psalms 30:11#
Tanach with Nikkud, via Sefaria.org
שְׁמַע־יְהוָה וְחָנֵּנִי יְהוָה הֱ⟨ZWJ⟩יֵה־עֹזֵר לִי׃
the Westminster Leningrad Codex as encoded by OpenScriptures
שְׁמַע־יְהוָה וְחָנֵּנִי יְהוָה הֱ∅יֵה־עֹזֵר לִי׃
The first copy carries a zero-width joiner (U+200D) that the other does not. The two encodings of the same manuscript part here, by one mark. Both descend from the Leningrad Codex, so neither outranks the other, and this record does not choose between two editions it cannot tell apart. Re-run the comparison with npm run verify:hebrew.
Psalm 30 in Hebrew Psalms 30:4#
Tanach with Nikkud, via Sefaria.org
…מִן־שְׁאוֹל נַפְשִׁי חִיִּיתַנִי מיורדי־[מִיָּרְדִי־]␣בוֹר׃
the Westminster Leningrad Codex as encoded by OpenScriptures
…מִן־שְׁאוֹל נַפְשִׁי חִיִּיתַנִי מיורדי␣[מִיָּרְדִי]־בוֹר׃
A maqaf (U+05BE) where the other has a space (U+0020). The two encodings of the same manuscript part here, by one mark. Both descend from the Leningrad Codex, so neither outranks the other, and this record does not choose between two editions it cannot tell apart. Re-run the comparison with npm run verify:hebrew.
Psalm 30 (JPS 1917) Psalm 30:7#
Wikisource, proofread against page scans (the copy that would be published)
Now I had said in my security: '␣I shall never be moved.'
Sefaria, from the Open Siddur Project transcription
Now I had said in my security: '∅I shall never be moved.'
The first copy carries a space (U+0020) that the other does not. The two digitisations part here, and a third copy, keyed separately by Mechon Mamre, sides against the reading that would be published. This record does not publish a line it would have had to repair. Re-run the comparison with npm run verify:jps.
O all-holy Trinity (Hapgood 1906) THE THIRD HOUR, page 43#
1906 first edition, Cornell University Library copy (the copy that would be published)
…upon us. O Lord, wash away our sins. O Master, pardon our transgression…
1922 second edition, Association Press
…upon us. O Lord, wash away our sins. © Master, pardon our transgression…
"O" (U+004F) where the other has a copyright sign (U+00A9). Two physical copies of the book, scanned and read separately, report different characters at the same place. A misreading in one is the likeliest cause, but correcting the second witness is how a second witness stops being one, so nothing is corrected and nothing is published. Re-run the comparison with npm run verify:scans.
The Lord's Prayer (Hapgood 1906) THE THIRD HOUR, page 44#
1906 first edition, Cornell University Library copy (the copy that would be published)
Our Father, who art in heaven. Hallowed be thy Name. Thy kingdo…
1922 second edition, Association Press
Our Father, who art in heaven, Hallowed be thy Name. Thy kingdo…
A full stop (U+002E) where the other has a comma (U+002C). Two physical copies of the book, scanned and read separately, report different characters at the same place. A misreading in one is the likeliest cause, but correcting the second witness is how a second witness stops being one, so nothing is corrected and nothing is published. Re-run the comparison with npm run verify:scans.

What those disagreements caught

One pass returned seventeenth-century manuscript spelling for an 1892 American prayer book, which is a wrong edition rather than a typo. Both passes garbled the same place name in an OCR-damaged line, which is exactly the text that should never ship. And one pair differed on whether a prayer for the dead reads “unto him” or “unto them”, a real variant that a single transcription would have resolved silently and confidently.

Every published text and its edition

Each entry has its own permanent link, so a specific text’s verification can be cited on its own rather than by pointing at this page and asking someone to search it.

Psalm 22 / 23 (Douay-Rheims)#
The Douay-Rheims Bible, 1899 American Edition, Psalms, 1899. Published 1899, before 1929, US public domain.
Located on the page between “Psalms Chapter 22 / Dominus regit me. / God's spiritual benefits to faithful souls.” and “Psalms Chapter 23”. Psalms Chapter 22, verses 22:1 to 22:6, in the plain text file drb2110.txt.
The scan this was read from
Psalm 129 / 130, De Profundis (Douay-Rheims)#
The Douay-Rheims Bible, 1899 American Edition, Psalms, 1899. Published 1899, before 1929, US public domain.
Located on the page between “Psalms Chapter 129 / De profundis. / A prayer of a sinner, trusting in the mercies of God.” and “Psalms Chapter 130”. Psalms Chapter 129, verses 129:1 to 129:8, in the plain text file drb2110.txt.
The scan this was read from
The Lord's Prayer (Douay-Rheims)#
The Douay-Rheims Bible, 1899 American Edition, 1899. Published 1899, before 1929, US public domain.
Located on the page between “6:8. Be not you therefore like to them for your Father knoweth what is needful for you, be” and “Lead us not into temptation... That is, suffer us not to be overcome by temptation.”. Matthew Chapter 6, verses 6:9-6:13, in the plain-text file 8300.txt (section headed "Matthew Chapter 6 / A continuation of the sermon on the mount.").
The scan this was read from
The Beatitudes (Douay-Rheims)#
The Douay-Rheims Bible, 1899 American Edition, 1899. Published 1899, before 1929, US public domain.
Located on the page between “5:2. And opening his mouth he taught them, saying:” and “5:13. You are the salt of the earth. But if the salt lose its savour, wherewith shall it b”. Matthew Chapter 5, verses 5:3-5:12, in the plain-text file 8300.txt (section headed "Matthew Chapter 5 / Christ's sermon upon the mount. The eight beatitudes.").
The scan this was read from
The Priestly Blessing (Singer 1904)#
The Authorised Daily Prayer Book, with a New Translation, seventh edition, 1904. Seventh edition, published 1904, before 1929, US public domain. The work first appeared in 1890 and the translator died in 1906, so it is also clear in life-plus-70 jurisdictions.
Located on the page between “At the repetition of the Amidah by the Reader, the following is introduced : —” and “At the Morning and Additional Services, and also on Fast Days at the Afternoon Service, sa”. English page of the Amidah, immediately after the Thanksgiving benediction ending "whose name is All-good, and unto whom it is becoming to give thanks.".
The scan this was read from
Psalm 121 (Singer 1904)#
The Authorised Daily Prayer Book, with a New Translation, seventh edition, 1904. Seventh edition, published 1904, before 1929, US public domain. The work first appeared in 1890 and the translator died in 1906, so it is also clear in life-plus-70 jurisdictions.
Located on the page between “Psalm cxxi. A Song of Degrees.” and “AA 2 [catchword at the foot of the page], then Psalm cxxii. A Song of Degrees ; of David.”. page 178 of the English, in the Songs of Degrees, immediately following Psalm cxx..
The scan this was read from
The Traveller's Prayer (Singer 1904)#
The Authorised Daily Prayer Book, with a New Translation, seventh edition, 1904. Seventh edition, published 1904, before 1929, US public domain. The work first appeared in 1890 and the translator died in 1906, so it is also clear in life-plus-70 jurisdictions.
Located on the page between “PRAYER TO BE SAID WHEN GOING ON A JOURNEY OR VOYAGE.” and “Then say, " And let the pleasantness" etc.. and Psalm xci., p. 294.”. page 310 of the volume, headed "Prayer to be said when going on a journey or voyage"; the preceding English matter ends with a rubric "Then say [Hebrew] p. 178.".
The scan this was read from
The Collect for Aid against All Perils#
The Book of Common Prayer, 1662, 1662. Published 1662, long before 1929, US public domain. UK Crown patent applies within the United Kingdom, so this edition is used here on the US basis only.
Located on the page between “1f The third Collect for Aid against all Perils.” and “^f Jn Quires and Places where they sing, here followeth the Ant heme.”. Order for Evening Prayer, third Collect; djvu.txt lines 15871-15877.
The scan this was read from
Psalm 23 (Coverdale)#
The Book of Common Prayer, 1662, 1662. Published 1662, long before 1929, US public domain. UK Crown patent applies within the United Kingdom, so this edition is used here on the US basis only.
Located on the page between “The xxiii. Psalm. Dominus regit me.” and “The xxiv. Psalm, Domini est terra.”. The Psalter, Psalm xxiii, page 355 and following; djvu.txt lines 33472-33518.
The scan this was read from
Psalm 121 (Coverdale)#
The Book of Common Prayer, 1662, 1662. Published 1662, long before 1929, US public domain. UK Crown patent applies within the United Kingdom, so this edition is used here on the US basis only.
Located on the page between “The Cxxi. Psalm Levari oculos.” and “The Cxxii. Psalm.”. The Psalter, Psalm cxxi, page 475; djvu.txt lines 42118-42155.
The scan this was read from
Psalm 130 (Coverdale)#
The Book of Common Prayer, 1662, 1662. Published 1662, long before 1929, US public domain. UK Crown patent applies within the United Kingdom, so this edition is used here on the US basis only.
Located on the page between “The Cxxx. Psalm. De profundi*.” and “The Cxxxi. Psalm. Domine turn est.”. The Psalter, Psalm cxxx; djvu.txt lines 42432-42456.
The scan this was read from
The General Thanksgiving#
The Book of Common Prayer, 1662, 1662. Published 1662, long before 1929, US public domain. UK Crown patent applies within the United Kingdom, so this edition is used here on the US basis only.
Located on the page between “A gentrall Thanksgiving.” and “^ For Rain”. Prayers and Thanksgivings vpon severall Occasions; djvu.txt lines 16790-16811.
The scan this was read from
A Prayer of St Chrysostom#
The Book of Common Prayer, 1662, 1662. Published 1662, long before 1929, US public domain. UK Crown patent applies within the United Kingdom, so this edition is used here on the US basis only.
Located on the page between “^f A Prayer of Saint Chrysostome.” and “% 2 Corinthians. 13. The grace of our Lord Iesus Christ, and the Love of God, and the Fell”. End of the Order for Evening Prayer; djvu.txt lines 15939-15949.
The scan this was read from
The Hail Mary (Baltimore Manual 1889)#
A Manual of Prayers for the Use of the Catholic Laity, 1889. Published 1889, before 1929, US public domain.
Located on the page between “lead us not into temptation : but deliver us from evil. Amen. The Hail Maty.” and “The Creed. CREDO in Deum, Pa-”. Ordinary morning prayers section, immediately after the Our Father and before the Creed.
The scan this was read from
The Memorare (Baltimore Manual 1889)#
A Manual of Prayers for the Use of the Catholic Laity, 1889. Published 1889, before 1929, US public domain.
Located on the page between “A Prayer ascribed to St. Bernard.” and “the Scapular OF THE CONFRATERNITY OF OUR LADY OF MT. CARMEL.”. pp. 381 to 382, running head on p. 382 is "Rosary of the Blessed Virgin Mary.".
The scan this was read from
Anima Christi (Baltimore Manual 1889)#
A Manual of Prayers for the Use of the Catholic Laity, 1889. Published 1889, before 1929, US public domain.
Located on the page between “A FAVORITE PRAYER OF ST. IGNATIUS LOYOLA. (i) Anima Christi.” and “At the Communion. LET it be now, O Lord, the effect of Thy mercy, that we who have been pr”. p. 103, in the Mass devotions after Communion.
The scan this was read from
Salve Regina (Baltimore Manual 1889)#
A Manual of Prayers for the Use of the Catholic Laity, 1889. Published 1889, before 1929, US public domain.
Located on the page between “IV. From First Vespers of Trinity Sunday co Advent.” and “Let us pray. ALMIGHTY, Everlasting God, who, by the co operation of the Holy Ghost, didst ”. pp. 79 to 80, the fourth of the seasonal Antiphons of the Blessed Virgin; running head on p. 80 is "Evening Prayers.".
The scan this was read from
The Memorare (Raccolta 1910)#
The Raccolta: or, Collection of Indulgenced Prayers, 1910. Published 1910, before 1929, US public domain.
Located on the page between “196. The Memorare. i. 300 Days. T.Q. ii. Plenary, once a month. I, II, III, IV. (See Instr” and “196 Pius IX, 11 Dec., 1846.”. No. 196, pp. 175-176.
The scan this was read from
Anima Christi (Raccolta 1910)#
The Raccolta: or, Collection of Indulgenced Prayers, 1910. Published 1910, before 1929, US public domain.
Located on the page between “95 Pius IX, January 9, 1854.” and “96. Prayer of St Alphonsus Liguori, to be said at a Visit to the most Holy Sacrament.”. No. 95, pp. 79-80.
The scan this was read from
El Padre Nuestro (Reina-Valera 1909)#
Santa Biblia, Reina-Valera, revision de 1909, 1909. Published 1909, before 1929, US public domain, and distributed as public domain by ebible.org. Later Reina-Valera revisions such as RVR1960 are in copyright and are not used here.
Located on the page between “No os hagáis, pues, semejantes á ellos; porque vuestro Padre sabe de qué cosas tenéis nece” and “Porque si perdonareis á los hombres sus ofensas, os perdonará también á vosotros vuestro P”. Mateo 6:9-13, chapter page MAT06.htm.
The scan this was read from
Salmo 23 (Reina-Valera 1909)#
Santa Biblia, Reina-Valera, revision de 1909, 1909. Published 1909, before 1929, US public domain, and distributed as public domain by ebible.org. Later Reina-Valera revisions such as RVR1960 are in copyright and are not used here.
Located on the page between “Salmo de David.” and “(end of chapter page; navigation link to Salmo 24 follows, no verse text of Salmo 24 on th”. Salmo 23, chapter page PSA023.htm.
The scan this was read from
Salmo 91 (Reina-Valera 1909)#
Santa Biblia, Reina-Valera, revision de 1909, 1909. Published 1909, before 1929, US public domain, and distributed as public domain by ebible.org. Later Reina-Valera revisions such as RVR1960 are in copyright and are not used here.
Located on the page between “91 (chapter number heading; this psalm carries no superscription in this edition)” and “(end of chapter page; navigation link to Salmo 92 follows, no verse text of Salmo 92 on th”. Salmo 91:1-16, chapter page PSA091.htm.
The scan this was read from
Las Bienaventuranzas (Reina-Valera 1909)#
Santa Biblia, Reina-Valera, revision de 1909, 1909. Published 1909, before 1929, US public domain, and distributed as public domain by ebible.org. Later Reina-Valera revisions such as RVR1960 are in copyright and are not used here.
Located on the page between “Y abriendo su boca, les enseñaba, diciendo:” and “Vosotros sois la sal de la tierra: y si la sal se desvaneciere ¿con qué será salada?”. Mateo 5:3-12, chapter page MAT05.htm.
The scan this was read from
Psalm 23 (JPS 1917)#
The Holy Scriptures According to the Masoretic Text (JPS), 1917. Published 1917, before 1929, US public domain.
Located on the page between “They shall come and shall declare His righteousness / Unto a people that shall be born, th” and “24 A Psalm of David. / The earth is the LORD's, and the fulness thereof;”. Psalms 23:1-6, Psalms book page (djvu pages 799-904).
The scan this was read from
Psalm 121 (JPS 1917)#
The Holy Scriptures According to the Masoretic Text (JPS), 1917. Published 1917, before 1929, US public domain.
Located on the page between “I am all peace; / But when I speak, they are for war. (end of Psalm 120, followed by the c” and “122 A Song of Ascents; of David. / I rejoiced when they said unto me:”. Psalms 121:1-8.
The scan this was read from
Psalm 130 (JPS 1917)#
The Holy Scriptures According to the Masoretic Text (JPS), 1917. Published 1917, before 1929, US public domain.
Located on the page between “'The blessing of the LORD be upon you; / We bless you in the name of the LORD.' (end of Ps” and “131 A Song of Ascents; of David. / Lord, my heart is not haughty, nor mine eyes lofty;”. Psalms 130:1-8.
The scan this was read from
The Priestly Blessing (JPS 1917)#
The Holy Scriptures According to the Masoretic Text (JPS), 1917. Published 1917, before 1929, US public domain.
Located on the page between “And the LORD spoke unto Moses, saying: 'Speak unto Aaron and unto his sons, saying: On thi” and “So shall they put My name upon the children of Israel, and I will bless them.'”. Numbers 6:24-26, on the Numbers book page.
The scan this was read from
The Shema, first paragraph#
The Holy Scriptures According to the Masoretic Text (JPS), 1917. Published 1917, before 1929, US public domain.
Located on the page between “Hear therefore, O Israel, and observe to do it; that it may be well with thee, and that ye” and “And it shall be, when the LORD thy God shall bring thee into the land which He swore unto ”. Deuteronomy 6:4-9, on the Deuteronomy book page.
The scan this was read from
Le Notre Pere (Segond 1910)#
La Sainte Bible, traduction Louis Segond, edition de 1910, 1910. Published 1910, before 1929, US public domain; the translator died in 1885. Later Segond revisions such as Segond 21 are in copyright and are not used here.
Located on the page between “6:8 Ne leur ressemblez pas; car votre Père sait de quoi vous avez besoin, avant que vous l” and “6:14 Si vous pardonnez aux hommes leurs offenses, votre Père céleste vous pardonnera aussi”. mat.htm, verse anchors an_6:9 through an_6:13.
The scan this was read from
Psaume 23 (Segond 1910)#
La Sainte Bible, traduction Louis Segond, edition de 1910, 1910. Published 1910, before 1929, US public domain; the translator died in 1885. Later Segond revisions such as Segond 21 are in copyright and are not used here.
Located on the page between “22:31 Quand elle viendra, elle annoncera sa justice, Elle annoncera son oeuvre au peuple n” and “24:1 Psaume de David. A l'Eternel la terre et ce qu'elle renferme, Le monde et ceux qui l'”. psa.htm, verse anchors an_23:1 through an_23:6.
The scan this was read from
Karaniya Metta Sutta (Fausboll 1881)#
The Sacred Books of the East, Vol. X: The Dhammapada and Sutta-Nipata, 1881. Published 1881, before 1929, US public domain, and both translators died before 1929.
Located on the page between “8. METTASUTTA. A peaceful mind and goodwill towards all beings are praised. — Text by Grim” and “Mettasutta is ended. 9. HEMAVATASUTTA. A dialogue between two Yakkhas on the qualities of ”. pp. 24-25, verses 1-10 (running numbers 142-151).
The scan this was read from
Ratana Sutta (Fausboll 1881)#
The Sacred Books of the East, Vol. X: The Dhammapada and Sutta-Nipata, 1881. Published 1881, before 1929, US public domain, and both translators died before 1929.
Located on the page between “Uragavagga is ended. II. K\3 LAVAGGA. 1. RATANASUTTA. For all beings salvation is only to ” and “Ratanasutta is ended. 2. Amagandhasutta. A bad mind and wicked deeds are what defiles a ma”. pp. 37-40, verses 1-17 (running numbers 221-237).
The scan this was read from
Salla Sutta, The Dart (Fausboll 1881)#
The Sacred Books of the East, Vol. X: The Dhammapada and Sutta-Nipata, 1881. Published 1881, before 1929, US public domain, and both translators died before 1929.
Located on the page between “Selasutta is ended. 8. SALLASUTTA. Life is short, all mortals are subject to death, but kn” and “Sallasutta is ended. 9. VASE77WASUTTA. A dispute arose between two young men, Bharadva^a a”. pp. 106-108, verses 1-20 (running numbers 574-593).
The scan this was read from

Hebrew, verified a different way

Two blind transcription passes work on a page a person reads and retypes. A vocalized Hebrew text is not that: it is a character encoding, and the failure it invites is not a misread word but a missing vowel point, which looks like nothing at all. So the Hebrew here is checked by comparing two independently maintained editions of the same text, character by character, vowel points included.

The Priestly Blessing (JPS 1917)#
Tanach with Nikkud, a digital edition of the Leningrad Codex, 2026. Tanach.us states that all biblical Hebrew text on its site may be viewed or copied without restriction, and Sefaria labels this version Public Domain. Retrieved and verified 2026-09-02.
Checked against the Westminster Leningrad Codex as encoded by OpenScriptures, a separate edition from a separate maintainer, released under Creative Commons Attribution 4.0. All 3 verses agreed codepoint for codepoint after Unicode NFC normalisation. The comparison removes only the Groves morpheme divider, the cantillation accents that a vowels-only edition does not carry, and the scribal paragraph markers; every vowel point, dagesh, maqaf and sof pasuq is compared as it stands.
Both editions descend from the Leningrad Codex, so this is evidence that neither encoding introduced an error. It is not independent manuscript witness, and tanach.us asks that its text not be called the Westminster Leningrad Codex, which is why the edition above is named as its publisher names it. Re-run the comparison with npm run verify:hebrew.
Psalm 100 (JPS 1917)#
Tanach with Nikkud, a digital edition of the Leningrad Codex, 2026. Tanach.us states that all biblical Hebrew text on its site may be viewed or copied without restriction, and Sefaria labels this version Public Domain. Retrieved and verified 2026-09-17.
Checked against the Westminster Leningrad Codex as encoded by OpenScriptures, a separate edition from a separate maintainer, released under Creative Commons Attribution 4.0. All 5 verses agreed codepoint for codepoint after Unicode NFC normalisation. The comparison removes only the Groves morpheme divider, the cantillation accents that a vowels-only edition does not carry, and the scribal paragraph markers; every vowel point, dagesh, maqaf and sof pasuq is compared as it stands.
Both editions descend from the Leningrad Codex, so this is evidence that neither encoding introduced an error. It is not independent manuscript witness, and tanach.us asks that its text not be called the Westminster Leningrad Codex, which is why the edition above is named as its publisher names it. Re-run the comparison with npm run verify:hebrew.
Psalm 103 (JPS 1917)#
Tanach with Nikkud, a digital edition of the Leningrad Codex, 2026. Tanach.us states that all biblical Hebrew text on its site may be viewed or copied without restriction, and Sefaria labels this version Public Domain. Retrieved and verified 2026-09-17.
Checked against the Westminster Leningrad Codex as encoded by OpenScriptures, a separate edition from a separate maintainer, released under Creative Commons Attribution 4.0. All 22 verses agreed codepoint for codepoint after Unicode NFC normalisation. The comparison removes only the Groves morpheme divider, the cantillation accents that a vowels-only edition does not carry, and the scribal paragraph markers; every vowel point, dagesh, maqaf and sof pasuq is compared as it stands.
Both editions descend from the Leningrad Codex, so this is evidence that neither encoding introduced an error. It is not independent manuscript witness, and tanach.us asks that its text not be called the Westminster Leningrad Codex, which is why the edition above is named as its publisher names it. Re-run the comparison with npm run verify:hebrew.
Psalm 113 (JPS 1917)#
Tanach with Nikkud, a digital edition of the Leningrad Codex, 2026. Tanach.us states that all biblical Hebrew text on its site may be viewed or copied without restriction, and Sefaria labels this version Public Domain. Retrieved and verified 2026-09-17.
Checked against the Westminster Leningrad Codex as encoded by OpenScriptures, a separate edition from a separate maintainer, released under Creative Commons Attribution 4.0. All 9 verses agreed codepoint for codepoint after Unicode NFC normalisation. The comparison removes only the Groves morpheme divider, the cantillation accents that a vowels-only edition does not carry, and the scribal paragraph markers; every vowel point, dagesh, maqaf and sof pasuq is compared as it stands.
Both editions descend from the Leningrad Codex, so this is evidence that neither encoding introduced an error. It is not independent manuscript witness, and tanach.us asks that its text not be called the Westminster Leningrad Codex, which is why the edition above is named as its publisher names it. Re-run the comparison with npm run verify:hebrew.
Psalm 121 (JPS 1917)#
Tanach with Nikkud, a digital edition of the Leningrad Codex, 2026. Tanach.us states that all biblical Hebrew text on its site may be viewed or copied without restriction, and Sefaria labels this version Public Domain. Retrieved and verified 2026-09-02.
Checked against the Westminster Leningrad Codex as encoded by OpenScriptures, a separate edition from a separate maintainer, released under Creative Commons Attribution 4.0. All 8 verses agreed codepoint for codepoint after Unicode NFC normalisation. The comparison removes only the Groves morpheme divider, the cantillation accents that a vowels-only edition does not carry, and the scribal paragraph markers; every vowel point, dagesh, maqaf and sof pasuq is compared as it stands.
Both editions descend from the Leningrad Codex, so this is evidence that neither encoding introduced an error. It is not independent manuscript witness, and tanach.us asks that its text not be called the Westminster Leningrad Codex, which is why the edition above is named as its publisher names it. Re-run the comparison with npm run verify:hebrew.
Psalm 121 (Singer 1904)#
Tanach with Nikkud, a digital edition of the Leningrad Codex, 2026. Tanach.us states that all biblical Hebrew text on its site may be viewed or copied without restriction, and Sefaria labels this version Public Domain. Retrieved and verified 2026-09-04.
Checked against the Westminster Leningrad Codex as encoded by OpenScriptures, a separate edition from a separate maintainer, released under Creative Commons Attribution 4.0. All 8 verses agreed codepoint for codepoint after Unicode NFC normalisation. The comparison removes only the Groves morpheme divider, the cantillation accents that a vowels-only edition does not carry, and the scribal paragraph markers; every vowel point, dagesh, maqaf and sof pasuq is compared as it stands.
Both editions descend from the Leningrad Codex, so this is evidence that neither encoding introduced an error. It is not independent manuscript witness, and tanach.us asks that its text not be called the Westminster Leningrad Codex, which is why the edition above is named as its publisher names it. Re-run the comparison with npm run verify:hebrew.
Psalm 126 (JPS 1917)#
Tanach with Nikkud, a digital edition of the Leningrad Codex, 2026. Tanach.us states that all biblical Hebrew text on its site may be viewed or copied without restriction, and Sefaria labels this version Public Domain. Retrieved and verified 2026-09-17.
Checked against the Westminster Leningrad Codex as encoded by OpenScriptures, a separate edition from a separate maintainer, released under Creative Commons Attribution 4.0. All 6 verses agreed codepoint for codepoint after Unicode NFC normalisation. The comparison removes only the Groves morpheme divider, the cantillation accents that a vowels-only edition does not carry, and the scribal paragraph markers; every vowel point, dagesh, maqaf and sof pasuq is compared as it stands.
Both editions descend from the Leningrad Codex, so this is evidence that neither encoding introduced an error. It is not independent manuscript witness, and tanach.us asks that its text not be called the Westminster Leningrad Codex, which is why the edition above is named as its publisher names it. Re-run the comparison with npm run verify:hebrew.
Psalm 128 (JPS 1917)#
Tanach with Nikkud, a digital edition of the Leningrad Codex, 2026. Tanach.us states that all biblical Hebrew text on its site may be viewed or copied without restriction, and Sefaria labels this version Public Domain. Retrieved and verified 2026-09-17.
Checked against the Westminster Leningrad Codex as encoded by OpenScriptures, a separate edition from a separate maintainer, released under Creative Commons Attribution 4.0. All 6 verses agreed codepoint for codepoint after Unicode NFC normalisation. The comparison removes only the Groves morpheme divider, the cantillation accents that a vowels-only edition does not carry, and the scribal paragraph markers; every vowel point, dagesh, maqaf and sof pasuq is compared as it stands.
Both editions descend from the Leningrad Codex, so this is evidence that neither encoding introduced an error. It is not independent manuscript witness, and tanach.us asks that its text not be called the Westminster Leningrad Codex, which is why the edition above is named as its publisher names it. Re-run the comparison with npm run verify:hebrew.
Psalm 13 (JPS 1917)#
Tanach with Nikkud, a digital edition of the Leningrad Codex, 2026. Tanach.us states that all biblical Hebrew text on its site may be viewed or copied without restriction, and Sefaria labels this version Public Domain. Retrieved and verified 2026-09-17.
Checked against the Westminster Leningrad Codex as encoded by OpenScriptures, a separate edition from a separate maintainer, released under Creative Commons Attribution 4.0. All 6 verses agreed codepoint for codepoint after Unicode NFC normalisation. The comparison removes only the Groves morpheme divider, the cantillation accents that a vowels-only edition does not carry, and the scribal paragraph markers; every vowel point, dagesh, maqaf and sof pasuq is compared as it stands.
Both editions descend from the Leningrad Codex, so this is evidence that neither encoding introduced an error. It is not independent manuscript witness, and tanach.us asks that its text not be called the Westminster Leningrad Codex, which is why the edition above is named as its publisher names it. Re-run the comparison with npm run verify:hebrew.
Psalm 130 (JPS 1917)#
Tanach with Nikkud, a digital edition of the Leningrad Codex, 2026. Tanach.us states that all biblical Hebrew text on its site may be viewed or copied without restriction, and Sefaria labels this version Public Domain. Retrieved and verified 2026-09-02.
Checked against the Westminster Leningrad Codex as encoded by OpenScriptures, a separate edition from a separate maintainer, released under Creative Commons Attribution 4.0. All 8 verses agreed codepoint for codepoint after Unicode NFC normalisation. The comparison removes only the Groves morpheme divider, the cantillation accents that a vowels-only edition does not carry, and the scribal paragraph markers; every vowel point, dagesh, maqaf and sof pasuq is compared as it stands.
Both editions descend from the Leningrad Codex, so this is evidence that neither encoding introduced an error. It is not independent manuscript witness, and tanach.us asks that its text not be called the Westminster Leningrad Codex, which is why the edition above is named as its publisher names it. Re-run the comparison with npm run verify:hebrew.
Psalm 150 (JPS 1917)#
Tanach with Nikkud, a digital edition of the Leningrad Codex, 2026. Tanach.us states that all biblical Hebrew text on its site may be viewed or copied without restriction, and Sefaria labels this version Public Domain. Retrieved and verified 2026-09-17.
Checked against the Westminster Leningrad Codex as encoded by OpenScriptures, a separate edition from a separate maintainer, released under Creative Commons Attribution 4.0. All 6 verses agreed codepoint for codepoint after Unicode NFC normalisation. The comparison removes only the Groves morpheme divider, the cantillation accents that a vowels-only edition does not carry, and the scribal paragraph markers; every vowel point, dagesh, maqaf and sof pasuq is compared as it stands.
Both editions descend from the Leningrad Codex, so this is evidence that neither encoding introduced an error. It is not independent manuscript witness, and tanach.us asks that its text not be called the Westminster Leningrad Codex, which is why the edition above is named as its publisher names it. Re-run the comparison with npm run verify:hebrew.
Psalm 16 (JPS 1917)#
Tanach with Nikkud, a digital edition of the Leningrad Codex, 2026. Tanach.us states that all biblical Hebrew text on its site may be viewed or copied without restriction, and Sefaria labels this version Public Domain. Retrieved and verified 2026-09-17.
Checked against the Westminster Leningrad Codex as encoded by OpenScriptures, a separate edition from a separate maintainer, released under Creative Commons Attribution 4.0. All 11 verses agreed codepoint for codepoint after Unicode NFC normalisation. The comparison removes only the Groves morpheme divider, the cantillation accents that a vowels-only edition does not carry, and the scribal paragraph markers; every vowel point, dagesh, maqaf and sof pasuq is compared as it stands.
Both editions descend from the Leningrad Codex, so this is evidence that neither encoding introduced an error. It is not independent manuscript witness, and tanach.us asks that its text not be called the Westminster Leningrad Codex, which is why the edition above is named as its publisher names it. Re-run the comparison with npm run verify:hebrew.
Psalm 20 (JPS 1917)#
Tanach with Nikkud, a digital edition of the Leningrad Codex, 2026. Tanach.us states that all biblical Hebrew text on its site may be viewed or copied without restriction, and Sefaria labels this version Public Domain. Retrieved and verified 2026-09-17.
Checked against the Westminster Leningrad Codex as encoded by OpenScriptures, a separate edition from a separate maintainer, released under Creative Commons Attribution 4.0. All 10 verses agreed codepoint for codepoint after Unicode NFC normalisation. The comparison removes only the Groves morpheme divider, the cantillation accents that a vowels-only edition does not carry, and the scribal paragraph markers; every vowel point, dagesh, maqaf and sof pasuq is compared as it stands.
Both editions descend from the Leningrad Codex, so this is evidence that neither encoding introduced an error. It is not independent manuscript witness, and tanach.us asks that its text not be called the Westminster Leningrad Codex, which is why the edition above is named as its publisher names it. Re-run the comparison with npm run verify:hebrew.
Psalm 23 (JPS 1917)#
Tanach with Nikkud, a digital edition of the Leningrad Codex, 2026. Tanach.us states that all biblical Hebrew text on its site may be viewed or copied without restriction, and Sefaria labels this version Public Domain. Retrieved and verified 2026-09-02.
Checked against the Westminster Leningrad Codex as encoded by OpenScriptures, a separate edition from a separate maintainer, released under Creative Commons Attribution 4.0. All 6 verses agreed codepoint for codepoint after Unicode NFC normalisation. The comparison removes only the Groves morpheme divider, the cantillation accents that a vowels-only edition does not carry, and the scribal paragraph markers; every vowel point, dagesh, maqaf and sof pasuq is compared as it stands.
Both editions descend from the Leningrad Codex, so this is evidence that neither encoding introduced an error. It is not independent manuscript witness, and tanach.us asks that its text not be called the Westminster Leningrad Codex, which is why the edition above is named as its publisher names it. Re-run the comparison with npm run verify:hebrew.
Psalm 27 (JPS 1917)#
Tanach with Nikkud, a digital edition of the Leningrad Codex, 2026. Tanach.us states that all biblical Hebrew text on its site may be viewed or copied without restriction, and Sefaria labels this version Public Domain. Retrieved and verified 2026-09-17.
Checked against the Westminster Leningrad Codex as encoded by OpenScriptures, a separate edition from a separate maintainer, released under Creative Commons Attribution 4.0. All 14 verses agreed codepoint for codepoint after Unicode NFC normalisation. The comparison removes only the Groves morpheme divider, the cantillation accents that a vowels-only edition does not carry, and the scribal paragraph markers; every vowel point, dagesh, maqaf and sof pasuq is compared as it stands.
Both editions descend from the Leningrad Codex, so this is evidence that neither encoding introduced an error. It is not independent manuscript witness, and tanach.us asks that its text not be called the Westminster Leningrad Codex, which is why the edition above is named as its publisher names it. Re-run the comparison with npm run verify:hebrew.
Psalm 46 (JPS 1917)#
Tanach with Nikkud, a digital edition of the Leningrad Codex, 2026. Tanach.us states that all biblical Hebrew text on its site may be viewed or copied without restriction, and Sefaria labels this version Public Domain. Retrieved and verified 2026-09-17.
Checked against the Westminster Leningrad Codex as encoded by OpenScriptures, a separate edition from a separate maintainer, released under Creative Commons Attribution 4.0. All 12 verses agreed codepoint for codepoint after Unicode NFC normalisation. The comparison removes only the Groves morpheme divider, the cantillation accents that a vowels-only edition does not carry, and the scribal paragraph markers; every vowel point, dagesh, maqaf and sof pasuq is compared as it stands.
Both editions descend from the Leningrad Codex, so this is evidence that neither encoding introduced an error. It is not independent manuscript witness, and tanach.us asks that its text not be called the Westminster Leningrad Codex, which is why the edition above is named as its publisher names it. Re-run the comparison with npm run verify:hebrew.
Psalm 6 (JPS 1917)#
Tanach with Nikkud, a digital edition of the Leningrad Codex, 2026. Tanach.us states that all biblical Hebrew text on its site may be viewed or copied without restriction, and Sefaria labels this version Public Domain. Retrieved and verified 2026-09-17.
Checked against the Westminster Leningrad Codex as encoded by OpenScriptures, a separate edition from a separate maintainer, released under Creative Commons Attribution 4.0. All 11 verses agreed codepoint for codepoint after Unicode NFC normalisation. The comparison removes only the Groves morpheme divider, the cantillation accents that a vowels-only edition does not carry, and the scribal paragraph markers; every vowel point, dagesh, maqaf and sof pasuq is compared as it stands.
Both editions descend from the Leningrad Codex, so this is evidence that neither encoding introduced an error. It is not independent manuscript witness, and tanach.us asks that its text not be called the Westminster Leningrad Codex, which is why the edition above is named as its publisher names it. Re-run the comparison with npm run verify:hebrew.
Psalm 67 (JPS 1917)#
Tanach with Nikkud, a digital edition of the Leningrad Codex, 2026. Tanach.us states that all biblical Hebrew text on its site may be viewed or copied without restriction, and Sefaria labels this version Public Domain. Retrieved and verified 2026-09-17.
Checked against the Westminster Leningrad Codex as encoded by OpenScriptures, a separate edition from a separate maintainer, released under Creative Commons Attribution 4.0. All 8 verses agreed codepoint for codepoint after Unicode NFC normalisation. The comparison removes only the Groves morpheme divider, the cantillation accents that a vowels-only edition does not carry, and the scribal paragraph markers; every vowel point, dagesh, maqaf and sof pasuq is compared as it stands.
Both editions descend from the Leningrad Codex, so this is evidence that neither encoding introduced an error. It is not independent manuscript witness, and tanach.us asks that its text not be called the Westminster Leningrad Codex, which is why the edition above is named as its publisher names it. Re-run the comparison with npm run verify:hebrew.
Psalm 91 (JPS 1917)#
Tanach with Nikkud, a digital edition of the Leningrad Codex, 2026. Tanach.us states that all biblical Hebrew text on its site may be viewed or copied without restriction, and Sefaria labels this version Public Domain. Retrieved and verified 2026-09-17.
Checked against the Westminster Leningrad Codex as encoded by OpenScriptures, a separate edition from a separate maintainer, released under Creative Commons Attribution 4.0. All 16 verses agreed codepoint for codepoint after Unicode NFC normalisation. The comparison removes only the Groves morpheme divider, the cantillation accents that a vowels-only edition does not carry, and the scribal paragraph markers; every vowel point, dagesh, maqaf and sof pasuq is compared as it stands.
Both editions descend from the Leningrad Codex, so this is evidence that neither encoding introduced an error. It is not independent manuscript witness, and tanach.us asks that its text not be called the Westminster Leningrad Codex, which is why the edition above is named as its publisher names it. Re-run the comparison with npm run verify:hebrew.
The Shema, first paragraph#
Tanach with Nikkud, a digital edition of the Leningrad Codex, 2026. Tanach.us states that all biblical Hebrew text on its site may be viewed or copied without restriction, and Sefaria labels this version Public Domain. Retrieved and verified 2026-09-01.
Checked against the Westminster Leningrad Codex as encoded by OpenScriptures, a separate edition from a separate maintainer, released under Creative Commons Attribution 4.0. All 6 verses agreed codepoint for codepoint after Unicode NFC normalisation. The comparison removes only the Groves morpheme divider, the cantillation accents that a vowels-only edition does not carry, and the scribal paragraph markers; every vowel point, dagesh, maqaf and sof pasuq is compared as it stands.
Both editions descend from the Leningrad Codex, so this is evidence that neither encoding introduced an error. It is not independent manuscript witness, and tanach.us asks that its text not be called the Westminster Leningrad Codex, which is why the edition above is named as its publisher names it. Re-run the comparison with npm run verify:hebrew.

Three of the psalms here contain a word that is written one way and read another, at Psalms 6:4, 100:3 and 126:4. The edition prints the written form unpointed and the read form after it in square brackets, and it is published that way. The second edition keeps the read form in a note, so the comparison lifts it out and checks both forms and does not drop the note as commentary.

Two Jewish entries have no Hebrew here: the Priestly Blessing and the Traveller’s Prayer, both from Singer’s prayer book. Their English is liturgical rather than biblical, a framing paragraph and a prayer for a journey, so no Masoretic edition contains them and there is nothing to fill them from. A third Singer entry, Psalm 121, does carry Hebrew, and it is a Masoretic edition rather than Singer’s own facing page. The entry says so: two columns from two publishers.

Singer’s Hebrew cannot come from the scan’s text layer at all. The library ran English character recognition over the whole book, so the Hebrew half of every page came out as noise, strings like S^h-Sk ninr^K where a line of prayer should be. It exists only as page images, and reading pointed Hebrew off an image is the one job here where two model passes agreeing would prove least, since both would be tempted to recite a liturgy they already know. That is the real obstacle to the Kaddish, the Amidah and Mi Sheberach, and it is a harder one than the phrase “not yet transcribed” made it sound.

A correction, since this page is where corrections belong. These three entries were cited as Singer 1890 until 16 September 2026. The scan they were read from says Seventh Edition, 5664 to 1904 on its title page, over a note to that edition dated January 1904. The work first appeared in 1890, which is how the wrong year got in: it is the date of the book, and this record cites the copy that was opened. The rights position is unchanged, and not one word of any text moved.

English psalms, from two separate digitisations

The psalms added on 17 September 2026 were not read by a model at all. Two projects keyed the Jewish Publication Society’s 1917 translation independently: Wikisource, proofread against page scans, and the Open Siddur Project, whose transcription Sefaria serves. A psalm is published only where the two agree character for character, verse by verse.

Psalm 100 (JPS 1917)#
12 printed lines, agreeing with the second digitisation in every verse. Open the published copy
Psalm 103 (JPS 1917)#
46 printed lines, agreeing with the second digitisation except as follows. The two copies differ at verse 13. A third, keyed separately by Mechon Mamre, agrees there with the reading published, letter and punctuation mark for mark. It lowers capitals, so in that verse the capitals rest on one copy. Open the published copy
Psalm 113 (JPS 1917)#
20 printed lines, agreeing with the second digitisation in every verse. Open the published copy
Psalm 126 (JPS 1917)#
15 printed lines, agreeing with the second digitisation in every verse. Open the published copy
Psalm 128 (JPS 1917)#
13 printed lines, agreeing with the second digitisation in every verse. Open the published copy
Psalm 13 (JPS 1917)#
13 printed lines, agreeing with the second digitisation in every verse. Open the published copy
Psalm 150 (JPS 1917)#
13 printed lines, agreeing with the second digitisation in every verse. Open the published copy
Psalm 16 (JPS 1917)#
24 printed lines, agreeing with the second digitisation in every verse. Open the published copy
Psalm 20 (JPS 1917)#
21 printed lines, agreeing with the second digitisation in every verse. Open the published copy
Psalm 27 (JPS 1917)#
38 printed lines, agreeing with the second digitisation except as follows. The two copies differ at verse 5. A third, keyed separately by Mechon Mamre, agrees there with the reading published, letter and punctuation mark for mark. It lowers capitals, so in that verse the capitals rest on one copy. Open the published copy
Psalm 46 (JPS 1917)#
24 printed lines, agreeing with the second digitisation in every verse. Open the published copy
Psalm 6 (JPS 1917)#
22 printed lines, agreeing with the second digitisation in every verse. Open the published copy
Psalm 67 (JPS 1917)#
16 printed lines, agreeing with the second digitisation except as follows. The two copies differ at verse 4. A third, keyed separately by Mechon Mamre, agrees there with the reading published, letter and punctuation mark for mark. It lowers capitals, so in that verse the capitals rest on one copy. Open the published copy
Psalm 91 (JPS 1917)#
34 printed lines, agreeing with the second digitisation in every verse. Open the published copy

Two differences were treated as typography and not as wording. Curly and straight quotation marks count as the same mark. And the capital that opens a printed line is not compared, because the second copy stores each verse as running text, has no line breaks, and often lowers that letter: it cannot testify about a capital that exists only because of a break it does not record. A capital anywhere else is compared as it stands. So the words, spelling and punctuation rest on two witnesses, and the line breaks rest on one.

Where the two copies split, the psalm is held unless something other than a guess settles it, and the entry above says what did. Three times the second copy carried a stray full stop after a verse, at Psalms 27:5, 67:4 and 103:13, and a third copy, keyed separately by Mechon Mamre, agreed with the reading published. That third copy lowers capitals, so it can say which reading has the stray mark and nothing about capitals. Psalm 30 is still held: there the space inside an opening quotation mark is in the copy that would be published, the other two agree against it, and this record does not publish a line it has had to repair. Re-run the comparison with npm run verify:jps.

Hebrew from the Mishnah

The Leningrad Codex is a Bible and does not contain the Mishnah, so a formula the Mishnah records cannot be checked against it. It is checked instead against a second vocalized Mishnah, keyed by a separate project, codepoint for codepoint with every vowel point compared, and published only where the two agree exactly.

Baruch Dayan HaEmet#
Mishnah, Torat Emet edition, Mishnah Berakhot 9:2. Sefaria labels this version Public Domain. The Mishnah itself is ancient and the pointing follows the standard printed editions. Checked against Hebrew Wikisource's separately keyed vocalized Mishnah. Retrieved and verified 2026-09-20.
Checked against Hebrew Wikisource, משנה ברכות ט ניקוד, keyed by a separate project. Both carry vowel points, and every letter and point agreed. Re-run the comparison with npm run verify:rabbinic.

A correction, dated as the others on this page are. From 20 September to 6 October 2026 this record listed Baruch Dayan HaEmet under the Hebrew section above, which describes a comparison with the Westminster Leningrad Codex, and so said a check had been made that never was: the Codex does not contain the Mishnah. The text itself was checked, against the second Mishnah described here, and not one word of it changes. The fault was in how this page decided which account to print beside an entry. It used to decide by script, and every Hebrew text before this one had had the Codex check; it now decides by the method each edition records about itself.

English read from two scanned copies

Eastern Orthodoxy had a full dossier here and not one text, because the one substantial public-domain English source, Isabel Hapgood’s service book of 1906, has never been keyed. It exists as page images with a machine-read text layer. So these were verified by comparing two different physical copies, scanned by different libraries and read by separate recognition runs, whose errors are independent of each other.

With the Saints Give Rest (Hapgood 1906)#
Service Book of the Holy Orthodox-Catholic Apostolic (Greco-Russian) Church, 1906. Published 1906, before 1929, US public domain, and the translator died in 1928. Retrieved and verified 2026-09-17.
Page 384, in burial of the dead (laymen), between “spirits, . . . See page 2,70.) The Collect Hymn {Konddk), in Tone III.” and “The Ikos. Thou only art immortal, who hast created and fashioned man. ”. Checked against the 1922 second edition, Association Press, which agrees at the same place, character for character.
Open the printed page in the scan
O Heavenly King (Hapgood 1906)#
Service Book of the Holy Orthodox-Catholic Apostolic (Greco-Russian) Church, 1906. Published 1906, before 1929, US public domain, and the translator died in 1928. Retrieved and verified 2026-09-17.
Page 43, in the third hour, between “d the blessing, we say : Amen. Glory to thee, our God ; glory to thee.” and “O Holy God, Holy Mighty, Holy Immortal One, have mercy upon us. ( Thri”. Checked against the 1922 second edition, Association Press, which agrees at the same place, character for character.
Open the printed page in the scan
Christ is Risen (Hapgood 1906)#
Service Book of the Holy Orthodox-Catholic Apostolic (Greco-Russian) Church, 1906. Published 1906, before 1929, US public domain, and the translator died in 1928. Retrieved and verified 2026-09-17.
Page 237, in the hours at easter, between “Blessed is our God always, now, and ever, and unto ages of ages. Amen.” and “( Thrice.) The Choir also singeth the Hymn: Christ is risen . . {Thric”. Checked against the 1922 second edition, Association Press, which agrees at the same place, character for character.
Open the printed page in the scan
The Trisagion (Hapgood 1906)#
Service Book of the Holy Orthodox-Catholic Apostolic (Greco-Russian) Church, 1906. Published 1906, before 1929, US public domain, and the translator died in 1928. Retrieved and verified 2026-09-17.
Page 43, in the third hour, between “ us, and cleanse us from every stain ; and save our souls, O Good One.” and “( Thrice, and three reverences. ) Glory to the Father, and to the Son,”. Checked against the 1922 second edition, Association Press, which agrees at the same place, character for character.
Open the printed page in the scan

Two things are normalised, in what is published as well as in the comparison. A word split by the end of a column is rejoined, and whether it keeps its hyphen is settled from how the same scan spells it elsewhere, never from a dictionary. And the space these books set before a colon or semicolon is dropped, because recognition reports it inconsistently, sometimes within one sentence. Every letter, capital and punctuation mark is compared as it stands.

The check earned its place on its first run. It refused the burial prayer beginning “O God of spirits” because the 1906 scan reads he {she) where the other reads he (she): an italic bracket misread as a brace, which a single recognition run would have published. That prayer is held for a second reason too. The book prints it under the rubric “the Priest, meanwhile, saith, secretly”, and a text a book assigns to a priest is one whose saying has conditions. Those wait for a reviewer however clean the scan is, which is also why the anointing prayer from the service for the sick is absent.

Where each passage sits is worked out and not eyeballed. These books print a running head and a page number on every page, and the scan’s text carries both, so the page is the number beside the nearest running head and the leaf comes from the scan’s own page index. The first three entries here had their page and leaf read off the page images by hand, and the method returns the same page and the same leaf for all three.

Three more prayers were tried on 17 September 2026 and are held, each for a reason worth stating. The Lord’s Prayer, because the 1906 copy ends its first clause with a full stop and the 1922 copy with a comma. “O all-holy Trinity”, because the 1922 scan reads © Master where the 1906 reads “O Master”, and a copyright sign is plainly a misreading, but correcting the second witness is how a second witness stops being one. And “Vouchsafe, O Lord”, because the two copies differ over whether the word is Name or name, and this book prints that prayer in a day form and a night form that are not the same. So of six prayers tried beyond the first two, one is published and three are held on one character each.

Hapgood is not consistent with herself. “O heavenly King” appears in six different punctuations inside the one 1906 volume, so each text here is cited to a page and a service rather than to the book, and the second copy had to agree at that same place. The second copy is the 1922 second edition and not another printing of the first, so agreement shows the wording stands in both; where she revised a passage the two differ and the text is simply not published. Nothing here can tell a revision from a misreading, and it does not try.

Arabic, and the first licensed edition here

Islam was blocked on this site until now, and the blocker was the English rather than the Arabic. The securely public-domain English Qur’an renderings are nineteenth-century orientalist work that no Muslim community uses devotionally, and printing one beside the Arabic would present it as the meaning of the prayer. So the Arabic is published and no English is. For a reciter the Arabic is the prayer.

These are the first texts here that are not in the public domain. Tanzil grants permission to copy and distribute verbatim copies of its Qur’an text provided the source is named and tanzil.net is linked, and forbids altering the text. All three conditions are met on every entry. Saying so is the point: a licence is not the public domain, and an edition that sits among public-domain ones without saying which it is would be the quiet kind of wrong this record exists to prevent.

Ad-Duha, the Morning Brightness#
The Tanzil Uthmani text. Tanzil Project (https://tanzil.net/), served by alquran.cloud.
All 12 lines agree with the Uthmani text served by the quran.com API, compared character by character after Unicode NFC normalisation. The comparison sets aside the byte-order mark, the tatweel, and the waqf, tajweed, sajdah and rub-el-hizb signs, each of which Tanzil itself offers as a download option and none of which is a letter of the text. Every haraka, the shadda, the superscript alef, alef wasla and the small waw and yeh were compared as they stand: those are orthography, not annotation.
One limit on the claim. The published text is Tanzil’s verbatim and carries the annotation marks, so the two editions were compared on the text under them and not on where a reciter pauses. Re-run the comparison with npm run verify:arabic.
Al-Falaq, the Daybreak#
The Tanzil Uthmani text. Tanzil Project (https://tanzil.net/), served by alquran.cloud.
All 6 lines agree with the Uthmani text served by the quran.com API, compared character by character after Unicode NFC normalisation. The comparison sets aside the byte-order mark, the tatweel, and the waqf, tajweed, sajdah and rub-el-hizb signs, each of which Tanzil itself offers as a download option and none of which is a letter of the text. Every haraka, the shadda, the superscript alef, alef wasla and the small waw and yeh were compared as they stand: those are orthography, not annotation.
One limit on the claim. The published text is Tanzil’s verbatim and carries the annotation marks, so the two editions were compared on the text under them and not on where a reciter pauses. Re-run the comparison with npm run verify:arabic.
Al-Fatiha, the Opening#
The Tanzil Uthmani text. Tanzil Project (https://tanzil.net/), served by alquran.cloud.
All 7 lines agree with the Uthmani text served by the quran.com API, compared character by character after Unicode NFC normalisation. The comparison sets aside the byte-order mark, the tatweel, and the waqf, tajweed, sajdah and rub-el-hizb signs, each of which Tanzil itself offers as a download option and none of which is a letter of the text. Every haraka, the shadda, the superscript alef, alef wasla and the small waw and yeh were compared as they stand: those are orthography, not annotation.
One limit on the claim. The published text is Tanzil’s verbatim and carries the annotation marks, so the two editions were compared on the text under them and not on where a reciter pauses. Re-run the comparison with npm run verify:arabic.
Al-Ikhlas, Sincerity#
The Tanzil Uthmani text. Tanzil Project (https://tanzil.net/), served by alquran.cloud.
All 5 lines agree with the Uthmani text served by the quran.com API, compared character by character after Unicode NFC normalisation. The comparison sets aside the byte-order mark, the tatweel, and the waqf, tajweed, sajdah and rub-el-hizb signs, each of which Tanzil itself offers as a download option and none of which is a letter of the text. Every haraka, the shadda, the superscript alef, alef wasla and the small waw and yeh were compared as they stand: those are orthography, not annotation.
One limit on the claim. The published text is Tanzil’s verbatim and carries the annotation marks, so the two editions were compared on the text under them and not on where a reciter pauses. Re-run the comparison with npm run verify:arabic.
The closing verses of Al-Baqarah#
The Tanzil Uthmani text. Tanzil Project (https://tanzil.net/), served by alquran.cloud.
All 2 lines agree with the Uthmani text served by the quran.com API, compared character by character after Unicode NFC normalisation. The comparison sets aside the byte-order mark, the tatweel, and the waqf, tajweed, sajdah and rub-el-hizb signs, each of which Tanzil itself offers as a download option and none of which is a letter of the text. Every haraka, the shadda, the superscript alef, alef wasla and the small waw and yeh were compared as they stand: those are orthography, not annotation.
One limit on the claim. The published text is Tanzil’s verbatim and carries the annotation marks, so the two editions were compared on the text under them and not on where a reciter pauses. Re-run the comparison with npm run verify:arabic.
An-Nas, Mankind#
The Tanzil Uthmani text. Tanzil Project (https://tanzil.net/), served by alquran.cloud.
All 7 lines agree with the Uthmani text served by the quran.com API, compared character by character after Unicode NFC normalisation. The comparison sets aside the byte-order mark, the tatweel, and the waqf, tajweed, sajdah and rub-el-hizb signs, each of which Tanzil itself offers as a download option and none of which is a letter of the text. Every haraka, the shadda, the superscript alef, alef wasla and the small waw and yeh were compared as they stand: those are orthography, not annotation.
One limit on the claim. The published text is Tanzil’s verbatim and carries the annotation marks, so the two editions were compared on the text under them and not on where a reciter pauses. Re-run the comparison with npm run verify:arabic.
The verse of marriage, Ar-Rum 21#
The Tanzil Uthmani text. Tanzil Project (https://tanzil.net/), served by alquran.cloud.
All 1 lines agree with the Uthmani text served by the quran.com API, compared character by character after Unicode NFC normalisation. The comparison sets aside the byte-order mark, the tatweel, and the waqf, tajweed, sajdah and rub-el-hizb signs, each of which Tanzil itself offers as a download option and none of which is a letter of the text. Every haraka, the shadda, the superscript alef, alef wasla and the small waw and yeh were compared as they stand: those are orthography, not annotation.
One limit on the claim. The published text is Tanzil’s verbatim and carries the annotation marks, so the two editions were compared on the text under them and not on where a reciter pauses. Re-run the comparison with npm run verify:arabic.
Ash-Sharh, the Opening of the Breast#
The Tanzil Uthmani text. Tanzil Project (https://tanzil.net/), served by alquran.cloud.
All 9 lines agree with the Uthmani text served by the quran.com API, compared character by character after Unicode NFC normalisation. The comparison sets aside the byte-order mark, the tatweel, and the waqf, tajweed, sajdah and rub-el-hizb signs, each of which Tanzil itself offers as a download option and none of which is a letter of the text. Every haraka, the shadda, the superscript alef, alef wasla and the small waw and yeh were compared as they stand: those are orthography, not annotation.
One limit on the claim. The published text is Tanzil’s verbatim and carries the annotation marks, so the two editions were compared on the text under them and not on where a reciter pauses. Re-run the comparison with npm run verify:arabic.
Ayat al-Kursi, the Verse of the Throne#
The Tanzil Uthmani text. Tanzil Project (https://tanzil.net/), served by alquran.cloud.
All 1 lines agree with the Uthmani text served by the quran.com API, compared character by character after Unicode NFC normalisation. The comparison sets aside the byte-order mark, the tatweel, and the waqf, tajweed, sajdah and rub-el-hizb signs, each of which Tanzil itself offers as a download option and none of which is a letter of the text. Every haraka, the shadda, the superscript alef, alef wasla and the small waw and yeh were compared as they stand: those are orthography, not annotation.
One limit on the claim. The published text is Tanzil’s verbatim and carries the annotation marks, so the two editions were compared on the text under them and not on where a reciter pauses. Re-run the comparison with npm run verify:arabic.
The prayer of Ayyub in illness#
The Tanzil Uthmani text. Tanzil Project (https://tanzil.net/), served by alquran.cloud.
All 1 lines agree with the Uthmani text served by the quran.com API, compared character by character after Unicode NFC normalisation. The comparison sets aside the byte-order mark, the tatweel, and the waqf, tajweed, sajdah and rub-el-hizb signs, each of which Tanzil itself offers as a download option and none of which is a letter of the text. Every haraka, the shadda, the superscript alef, alef wasla and the small waw and yeh were compared as they stand: those are orthography, not annotation.
One limit on the claim. The published text is Tanzil’s verbatim and carries the annotation marks, so the two editions were compared on the text under them and not on where a reciter pauses. Re-run the comparison with npm run verify:arabic.
The prayer of Ibrahim for his family#
The Tanzil Uthmani text. Tanzil Project (https://tanzil.net/), served by alquran.cloud.
All 2 lines agree with the Uthmani text served by the quran.com API, compared character by character after Unicode NFC normalisation. The comparison sets aside the byte-order mark, the tatweel, and the waqf, tajweed, sajdah and rub-el-hizb signs, each of which Tanzil itself offers as a download option and none of which is a letter of the text. Every haraka, the shadda, the superscript alef, alef wasla and the small waw and yeh were compared as they stand: those are orthography, not annotation.
One limit on the claim. The published text is Tanzil’s verbatim and carries the annotation marks, so the two editions were compared on the text under them and not on where a reciter pauses. Re-run the comparison with npm run verify:arabic.
The prayer of Musa before a hard task#
The Tanzil Uthmani text. Tanzil Project (https://tanzil.net/), served by alquran.cloud.
All 4 lines agree with the Uthmani text served by the quran.com API, compared character by character after Unicode NFC normalisation. The comparison sets aside the byte-order mark, the tatweel, and the waqf, tajweed, sajdah and rub-el-hizb signs, each of which Tanzil itself offers as a download option and none of which is a letter of the text. Every haraka, the shadda, the superscript alef, alef wasla and the small waw and yeh were compared as they stand: those are orthography, not annotation.
One limit on the claim. The published text is Tanzil’s verbatim and carries the annotation marks, so the two editions were compared on the text under them and not on where a reciter pauses. Re-run the comparison with npm run verify:arabic.
The prayer of Yunus in distress#
The Tanzil Uthmani text. Tanzil Project (https://tanzil.net/), served by alquran.cloud.
All 1 lines agree with the Uthmani text served by the quran.com API, compared character by character after Unicode NFC normalisation. The comparison sets aside the byte-order mark, the tatweel, and the waqf, tajweed, sajdah and rub-el-hizb signs, each of which Tanzil itself offers as a download option and none of which is a letter of the text. Every haraka, the shadda, the superscript alef, alef wasla and the small waw and yeh were compared as they stand: those are orthography, not annotation.
One limit on the claim. The published text is Tanzil’s verbatim and carries the annotation marks, so the two editions were compared on the text under them and not on where a reciter pauses. Re-run the comparison with npm run verify:arabic.
The prayer of Zakariyya for a child#
The Tanzil Uthmani text. Tanzil Project (https://tanzil.net/), served by alquran.cloud.
All 1 lines agree with the Uthmani text served by the quran.com API, compared character by character after Unicode NFC normalisation. The comparison sets aside the byte-order mark, the tatweel, and the waqf, tajweed, sajdah and rub-el-hizb signs, each of which Tanzil itself offers as a download option and none of which is a letter of the text. Every haraka, the shadda, the superscript alef, alef wasla and the small waw and yeh were compared as they stand: those are orthography, not annotation.
One limit on the claim. The published text is Tanzil’s verbatim and carries the annotation marks, so the two editions were compared on the text under them and not on where a reciter pauses. Re-run the comparison with npm run verify:arabic.
Inna lillahi, the verse said at a loss#
The Tanzil Uthmani text. Tanzil Project (https://tanzil.net/), served by alquran.cloud.
All 1 lines agree with the Uthmani text served by the quran.com API, compared character by character after Unicode NFC normalisation. The comparison sets aside the byte-order mark, the tatweel, and the waqf, tajweed, sajdah and rub-el-hizb signs, each of which Tanzil itself offers as a download option and none of which is a letter of the text. Every haraka, the shadda, the superscript alef, alef wasla and the small waw and yeh were compared as they stand: those are orthography, not annotation.
One limit on the claim. The published text is Tanzil’s verbatim and carries the annotation marks, so the two editions were compared on the text under them and not on where a reciter pauses. Re-run the comparison with npm run verify:arabic.
The prayer for spouse and children#
The Tanzil Uthmani text. Tanzil Project (https://tanzil.net/), served by alquran.cloud.
All 1 lines agree with the Uthmani text served by the quran.com API, compared character by character after Unicode NFC normalisation. The comparison sets aside the byte-order mark, the tatweel, and the waqf, tajweed, sajdah and rub-el-hizb signs, each of which Tanzil itself offers as a download option and none of which is a letter of the text. Every haraka, the shadda, the superscript alef, alef wasla and the small waw and yeh were compared as they stand: those are orthography, not annotation.
One limit on the claim. The published text is Tanzil’s verbatim and carries the annotation marks, so the two editions were compared on the text under them and not on where a reciter pauses. Re-run the comparison with npm run verify:arabic.
Rabbana atina, for good in both worlds#
The Tanzil Uthmani text. Tanzil Project (https://tanzil.net/), served by alquran.cloud.
All 1 lines agree with the Uthmani text served by the quran.com API, compared character by character after Unicode NFC normalisation. The comparison sets aside the byte-order mark, the tatweel, and the waqf, tajweed, sajdah and rub-el-hizb signs, each of which Tanzil itself offers as a download option and none of which is a letter of the text. Every haraka, the shadda, the superscript alef, alef wasla and the small waw and yeh were compared as they stand: those are orthography, not annotation.
One limit on the claim. The published text is Tanzil’s verbatim and carries the annotation marks, so the two editions were compared on the text under them and not on where a reciter pauses. Re-run the comparison with npm run verify:arabic.

One encoding difference is folded, and it is the only one. Where a hamza sits on the line with no letter to carry it, as in the word for the hereafter, Tanzil writes the letter itself and the other service writes a joining stroke with a hamza mark above it, which draws the same thing. It turned up at 2:201, 30:21 and 93:4, each time at exactly that point and nowhere else. A hamza above any real letter is a different spelling and is still compared as it stands.

Several of these are a single verse in which someone prays, and the verse is given whole, including the words that introduce the prayer. Tanzil’s text is used on condition that it is not altered, and a verse cut down to its best-known phrase is altered.

What this record does not claim

That two transcriptions agree says the wording matches the edition we cite. It says nothing about whether a text is the right one to offer a particular person at a particular moment. That judgment needs someone credentialed in the tradition, and nobody has signed these. How review works here, and how rights are established.

Content last changed 6 October 2026. Dates come from this page’s own revision history, not from the last time the site was deployed.