About StudyTools
Free Bible study tools that run entirely in your browser — no accounts, no tracking, no server.
What this is
StudyTools is a collection of small, dependency-free Bible study apps. Everything is static HTML, CSS and JavaScript, so the whole site can be served from any static host with no build step and no backend.
Between them they read the Bible in three modern public-domain translations, search it word by word, write notes against a verse, carry a plan through the whole text a day at a time, keep the church's year from the Book of Common Prayer (1928), and open any passage in the Greek or Hebrew behind it.
Notes, journals, memory decks, reading-plan ticks and study progress are saved in your browser's local storage and never leave your device.
It can be installed on a phone or a laptop and then used with no connection at all: a service worker keeps the pages you have visited, so the site opens on a plane. And because there is no account to hang your reading on, Your Data gathers everything it has saved in one place — a backup file to keep, or a short code to carry to another device, where the two copies are added together rather than one overwriting the other.
The companion site is Witness, a cited term base for interfaith translation: how each tradition renders a concept, and what each rendering loses. It runs with no server too.
The Bible texts
Seven translations ship with the site, all in the public domain, which is why they can be bundled, copied, printed and shared freely. Three are modern and carry the Apocrypha: the World English Bible (Updated), the modernized King James Version and the modernized Revised Version. Two keep to the sixty-six book canon: the American Standard Version (1901) and Young's Literal Translation (1862), both brought into present-day English by this project like the Revised Version. One is Jewish: the Jewish Publication Society's Tanakh of 1917, modernized, for the Old Testament — the same reading the Parallel Passages tool sets beside a Christian one. One is Catholic and older: the Douay-Rheims of 1752, which reads the Latin Vulgate and numbers it the Vulgate's way (see below).
The Parallel Passages tool quotes the Qur'an, the Jewish Tanakh and the Book of Mormon beside the Bible on the same figures and events. Only the passages it compares are bundled here — the full texts are deliberately not part of this site — and each is quoted from its own public-domain edition, modernized the same way as the Bibles above.
Note: the three modern texts are genuinely independent translations, not editions of one another. Measured against each other, under 2% of verses are word-for-word identical between WEBU and RVM, and about 12% between KJVM and RVM. All three of those carry the Apocrypha, though they present it differently: WEBU has the Greek Esther and Daniel, Psalm 151 and 3–4 Maccabees, while KJVM and RVM keep the additions to Daniel as separate books (Susanna, Bel and the Dragon, the Prayer of Azariah).
| Translation | Books | Public domain because |
|---|---|---|
| World English Bible, Updated WEBU |
81 with the Apocrypha |
Not copyrighted. Published by eBible.org (Michael Paul Johnson). “World English Bible” is a trademark of eBible.org. |
| King James Version, Modernized KJVM |
81 with the Apocrypha |
The classic King James text brought to present-day English, from the Abrahamic Library kjv-bible work, and finished by this site's own rule-based pass — the source left 103 verses with a thee or a ye beside a you, and Holy Ghost throughout. KJVM is this site's own label for the edition — it is not the Modern King James Version, which is a different, copyrighted translation. |
| Revised Version, Modernized RVM |
80 with the Apocrypha |
The Revised Version of 1895 — public domain, from eBible.org — brought into present-day English by this project. scripts/modernize.py updates the second-person pronouns and verb forms (thou knowest → you know, goeth → goes) and leaves the wording otherwise untouched. RVM is this site's own label for the edition. |
| American Standard Version, Modernized ASVM |
66 canon only |
The 1901 revision that the Revised Standard Version and the New American Standard Bible descend from — formal equivalence, still widely quoted. Text from scrollmapper/bible_databases, brought into present-day English by this project's scripts/modernize.py pass like the Revised Version. |
| Young's Literal Translation, Modernized YLTM |
66 canon only |
Robert Young's 1862 rendering, deliberately mechanical — it keeps the tense and the word order of the original, which is what makes it useful beside the interlinear. Text from the same scrollmapper collection as the American Standard Version, modernized by the same pass. |
| Douay-Rheims, Modernized DRCM · Vulgate numbering |
73 with the deuterocanon |
The Catholic English Bible of 1582–1610 in Bishop Challoner's revision of 1752, translated from the Latin Vulgate, so it reads differently as well as numbering differently — it has verses the critical text dropped (Acts 8:37, the Comma Johanneum). Text from scrollmapper/bible_databases, modernized by this project's pass. Read on its own, never placed beside the others. |
| JPS Tanakh, Modernized JPSM |
39 Hebrew Bible |
The Jewish Publication Society's 1917 translation of the Hebrew Bible, brought into present-day English by this project's rule-based pass like the Revised Version. The only translation here made from within the Jewish tradition rather than a Christian one. |
On numbering
The reader, Study a Passage and the verse panel keep every table that is keyed to the Hebrew numbering — cross-references, commentary, the interlinear, the Prayer Book's offices — and turn a Douay-Rheims reference into it before looking anything up. Its Psalm 22 shows what the commentators said about Psalm 23. The six Masoretic-numbered translations use the same chapter and verse numbering, so a reference lines up exactly between them. The Douay-Rheims does not: it keeps the Vulgate's, where Psalm 22 is the Hebrew Psalm 23, where Psalms 9 and 10 are one psalm and 114 and 115 are another, and where Daniel and Esther carry their additions as chapters. It is therefore read on its own — the reader says so on every chapter, names the Hebrew psalm you are actually in, and turns a shared link like Psalm 23 into the psalm the Douay-Rheims calls 22 — and it is kept out of the Verse Matrix, the verse panel and the study page, so that nothing is placed beside the wrong verse. That is the whole reason it is offered this way rather than dropped.
The World English Bible is not copyrighted; “World English Bible” is a trademark of eBible.org, and a modified text should not carry that name. Its full notice is at eBible.org.
The per-book text under data/bible/ is built from these sources by the scripts in scripts/.
Other sources
The study tools draw on these public-domain and openly licensed datasets:
- Cross-references — openbible.info · CC BY 4.0
- Atlas places and their verses — the OpenBible.info Bible Geocoding dataset, the same publisher as the cross-references · CC BY 4.0
- Atlas ground and coastlines — Natural Earth 1:50m shaded relief, coastlines, rivers, lakes and borders, and 1:110m for the world view · public domain
- Commentary — Jamieson, Fausset & Brown (1871), John Calvin (1550s) and F. B. Meyer, via the HelloAO Bible API · public domain
- Bible dictionaries — Easton (1897), Smith (1863), Hastings (1909) and Hitchcock's Bible Names (1869), via the NEUU bible-dictionary-dataset · CC BY 4.0
- Dictionary of words: Webster's 1828 — a transcription of the 1828, shipped whole at 60,968 entries with the words the King James Version uses marked and shown first (that word list taken from the KJV at eBible.org), which is what defines a word like lasciviousness. 2,004 rows of the transcription are its own error pages and are left out · public domain
- Notes in the dictionary — 121 entries written for this site where the five leave a gap a reader meets often: the words the transcription of the 1828 has a failure for, and the King James' own forms, compounds and spellings. Each one quotes the verse it is about, and
check-dictionary.jsrequires that verse to exist and to contain the word · this project's own writing - Topical Bible — Nave's (1897) and Torrey's New Topical Textbook (1897) · public domain
- Interlinear — Greek New Testament from OpenGNT · CC BY-SA 4.0; Hebrew Old Testament from the Open Scriptures Hebrew Bible with STEPBible TBESH glosses · CC BY 4.0
- Greek and Hebrew vocabulary — OpenGNT and the Open Scriptures Hebrew Bible · CC BY 4.0
- Lexicon and glosses — STEPBible TBESG / TBESH · CC BY 4.0
- Daily office and Sunday readings — Book of Common Prayer (1928), Episcopal Church, transcribed by vovchykbratyk/BCP_1928 · public domain
- Optional local model — WebLLM (Apache-2.0) running Llama 3.2 3B Instruct (Meta, Llama 3.2 Community Licence), downloaded to the reader's own device · off by default
- Creeds and catechisms — Apostles', Nicene and Athanasian; Heidelberg and Westminster Shorter · public domain
How the data is built and checked
The texts and tables the tools use are not typed in by hand. Each is built from its source by a script in scripts/, and the parts a reader would notice getting wrong are checked by scripts of their own, which anyone can run again:
- The church year — Easter by the usual computus, every season derived from it, every date from 1900 to 2100 confirmed to have both a morning and an evening office, and every Sunday named as the Prayer Book names it · check-liturgical.js
- The readings — the Prayer Book's printed references, alternatives and all, read back to book, chapter and verse · check-refs.js
- The reading plans — every chapter of a plan read exactly once, in order, with no empty day and no day more than one chapter longer than another · check-plan.js
- Search — the word index compared against the text it indexes, book by book and word by word, with the two tokenisers pinned to one another · check-search.js
- Notes — the store rules, both orderings and the Markdown export · check-notes.js
- Devotional verses — every quoted verse re-checked against the bundled text, so the site cannot print a misquotation · verify-devotional.py
- Concordance — the Strong's index re-checked against the interlinear it was built from, word by word: every verse a word is listed in has to contain it, and every verse it appears in has to be listed · check-concordance.js
- Modernized translations — each one re-checked against the text it was made from: the structure verse for verse, the archaic forms, and every word, so a rule cannot invent one · verify-modernized.py
- Commentary — the verses and introductions it is indexed to, and the books it has nothing for · check-commentary.js
- Ask the site — every question answered out of the files above, and the links each answer offers · check-answer.js
- The dictionary — five merged sources against the index that lists them: every definition able to cite one, no markup or mis-read characters left in the text, the abridgment within its length cap, and a King James word the Bible dictionaries have nothing on (lasciviousness) still defined · check-dictionary.js
- The atlas — the box the basemap was cropped to against the box the page draws it at, the raster's own shape against the degrees it claims, the ground against the window the map opens on, the name sizes in the stylesheet against the ones the collision boxes are measured with, no name drawn twice, and every verse a place is named in against the text in the reader · check-atlas.js
- The pages — what each one loads and says about itself, including a stylesheet and its own button styling · check-pages.js
- The offline copy — the service worker driven against a stubbed network: pages and scripts fetched past the browser's cache so a deploy is not left invisible for ten minutes, data revalidated, everything falling back to the cache with no connection, and every file the shell installs actually shipped · check-sw.js
Note: the commentary is public domain and uneven, and the panel says so where it is. The Gospels carry the church fathers as well: Aquinas' Catena Aurea, in the Oxford translation of 1842, on Matthew and Mark verse by verse, each remark attributed to the Father who made it. (CCEL holds no text volume for Luke or John, so those two have none.) The three main works cover the sixty-six; the deuterocanonical books are Haydock's — his Catholic Bible Commentary (1859) goes through Tobit, Judith, the Greek Esther, Wisdom, Sirach, Baruch, the Epistle of Jeremiah, Susanna, Bel and the Dragon and the Maccabees verse by verse, and his introduction to each of them is shown as about this book. Six books are outside his range, and for four of them — 1 and 2 Esdras, 3 and 4 Maccabees — the panel shows the article on the book from Hastings' Dictionary of the Bible (1909), which this site already bundles, under a line saying that is what it is. The Prayer of Manasses is covered by R. H. Charles (1913) — his introduction to it, transcribed from a scan of the edition, and disclosed in the panel as a transcription rather than commentary. One book has nothing at all: Psalm 151, which no public-domain work here treats as a book, and the panel says so. Because these commentators attach a remark to the verse that opens a passage, most individual verses have none — that is the shape of the material, not a fault in the reader. They also print front matter before most chapters, and that is carried too, as the chapter's introduction: 1,178 of them, which is where Jamieson, Fausset & Brown comment on Genesis 1:1, and the whole of what they wrote on the Song of Solomon. The three works are Protestant in range, so no deuterocanonical book has any commentary at all.
Where a check has found something wrong, the fix went into the data or the arithmetic rather than into a footnote: an Epiphany taking its year from Christmas, a Sunday Lectionary three weeks out because a year has fewer Sundays after the Epiphany than the tables allow for, and a lexicon lookup that left everyday Hebrew words blank.
Privacy
Nothing is uploaded. There are no accounts, no cookies, no analytics and no server to talk to. Everything you type — verse notes and highlights, journal entries, memory decks, reading-plan ticks, saved ideas — stays in your browser's local storage, and clearing your browser data removes it.
One thing can leave your device, and only if you ask for it. The reader can offer a language model to ask about a verse, and before you ask, the panel says which model it is and where the question will go. The default is a small one that runs on your own machine through WebGPU, so nothing you read or type is sent anywhere — but the model itself is downloaded once, from Hugging Face (via WebLLM), and kept in your browser's cache until you clear it. It is off until you turn it on, it tells you the download size before it starts, and it cannot verify anything it says.
Two other choices do send the question and the passage off this device, and the panel says so where you pick them. One is a model you run yourself — LM Studio, Ollama or vLLM — which is reached at localhost and never leaves your computer. The other is a provider you already have an account with: OpenRouter, Groq, Cerebras, Together, Fireworks or OpenAI. That one receives your question, the passage, and the API key you paste in, which is kept in this browser beside your notes and sent to nobody but that provider. This site never sees the question, the key, or the bill. There are no analytics anywhere in any of it, and no account here to hold anything: the only thing that ever leaves your device is what you deliberately ask to leave.
One more switch can send something. Ask the Site reads a question with rules — is there a reference in it, a Strong's number, a word the dictionaries know — and those rules are blind to some phrasings. If you turn on its assist, which is off by default and needs a key of your own, such a question is sent to the model you named so it can be asked one thing: which of the subjects this site can already answer is the question about. The question is all that leaves, no key of this site's is involved, and the answer you then read is still this site's own lookups on that word, cited as always. Questions the rules already understand are never sent anywhere.
Following the link to Witness, the companion site, takes you off this one — that site will see the visit, as any website does. Nothing here sends it anything.