The whole canon, queryable.
Bible CLI is an offline study toolkit for AI agents and humans: four public-domain English translations, the tagged Hebrew and Greek originals with full morphology, lexicons, 345,000 cross-references, and a computed map of where the New Testament quotes the Old. Every command speaks JSON.
Install
$ npm install -g @divine-creative-ministries/bible-cliThat is the whole install. The scripture databases (~90 MB, checksummed) download automatically the first time a command needs them. Or try it with zero install:
$ npx @divine-creative-ministries/bible-cli passage "John 3:16"
Data lives in ~/.bible-cli (override with
BIBLE_CLI_DATA). To manage downloads manually, set
BIBLE_CLI_NO_AUTO_DOWNLOAD=1 and run
bible db download. The optional Septuagint layer —
which powers bible quotes — is a separate artifact under
its own license: bible db download-lxx.
Requires Node ≥ 20.
Quick start
$ bible passage "John 3:16"
[BSB] John
3:16 For God so loved the world that He gave His one and only Son,
that everyone who believes in Him shall not perish but have
eternal life.$ bible search "living water" --book nt
$ bible search "covenant" --book Genesis --count
23 matching verses$ bible compare "Rom 8:1" -t all
Romans 8:1
ASV There is therefore now no condemnation to them that are in Christ Jesus.
BSB Therefore, there is now no condemnation for those who are in Christ Jesus.
KJV There is therefore now no condemnation to them which are in Christ Jesus,
who walk not after the flesh, but after the Spirit.
WEB There is therefore now no condemnation to those who are in Christ Jesus...Where translations diverge — as KJV does above, carrying a longer reading from its Greek source — an interpretive or textual decision is hiding. That is the signal to go one level down.
References are forgiving: John 3:16-18,
jn 3 16, 1jn2:5, Psalm 23,
Gen 1:1-2:3, and psa 51:title all work.
Scope flags accept books, ranges, and testaments:
--book Isaiah, --book Gen-Deu,
--book ot.
Original languages
$ bible interlinear "John 3:16"
Οὕτως Houtōs G3779 Adv so
γὰρ gar G1063 Conj For
ἠγάπησεν ēgapēsen G0025 V-AIA-3S loved
ὁ ho G3588 Art-NMS -
Θεὸς Theos G2316 N-NMS God
τὸν ton G3588 Art-AMS the
κόσμον kosmon G2889 N-AMS world$ bible word chesed
H2617 חֶ֫סֶד (chesed) — goodness, kindness
occurrences: 247 in 241 verses | top books: Psalms (127), 2 Samuel (12) ...
gloss range: covenant loyalty ×168; loyalty ×41; favor ×3 ...
[bdb] full Brown-Driver-Briggs entry follows$ bible grep-morph --stem niphal --tense participle --book Isaiah --count
98 matching words
$ bible original "John 1:18" --edition tr # Textus Receptus wording
$ bible lemma H2617 --book Psalms # every occurrence, in scope
The tagged text carries 469,000 Hebrew and Aramaic morphemes and
146,000 Greek words — each with parsing, disambiguated Strong's
numbers, and a context-sensitive gloss. Greek variant readings are
edition-tagged (NA, SBL, Textus Receptus, Byzantine), so
--edition renders real edition wording.
Pattern analysis
Start any study with survey — one call returning the
corpus's own structure for a topic before any interpretation:
$ bible survey chesed
H2617 חֶ֫סֶד (chesed) — goodness, kindness
occurrences: 247 in 241 verses | OT 247 / NT 0
distribution: Psalms 127, 2 Samuel 12, Proverbs 11, Genesis 11 ...
co-occurring distinctive words:
H5769 עוֹלָם "forever" — 57 shared verses
quotes is a computed map of where the New Testament
quotes, alludes to, and echoes the Old — built by matching the Greek
NT against the Septuagint, in labeled confidence tiers, at both
verbatim and lemma level:
$ bible quotes "Psalm 110:1"
Matthew 22:44 ⇐ Psalms 110:1 quotation 14w
Mark 12:36 ⇐ Psalms 110:1 quotation 14w
Hebrews 1:13 ⇐ Psalms 110:1 quotation 14w
Acts 2:34 ⇐ Psalms 110:1 quotation 10w ... 7 NT citations
$ bible quotes "Rev 1:7" --tier echo
Revelation 1:7 ⇐ Psalms 72:17 quotation 5w
Revelation 1:7 ⇐ Genesis 12:3 allusion 5w≈ lemma-level match
Revelation 1:7 ⇐ Daniel 7:13 echo 3w≈ "coming with the clouds"$ bible xref "Isa 53:5" --text # 345k vote-ranked cross-references
$ bible freq --strongs H1285 --by-book # where does "covenant" concentrate?
$ bible cooccur --strongs H2617 --strongs H1285
$ bible similar "Isa 53:3-7" # shared distinctive vocabulary
$ bible name Zechariah # which of the ~30 Zechariahs?For AI agents
Every command takes --json, errors are machine-readable
with did-you-mean suggestions, output is limit-disciplined with
count-only modes, and capabilities are discoverable at runtime
(bible books, translations,
morph-codes, licenses).
Install the study methodology
$ bible agent-setup claude # writes .claude/skills/bible-study/SKILL.md
$ bible agent-setup generic # appends to AGENTS.md (Codex, opencode, ...)
The methodology is a text-first protocol: evidence scaled to
the claim (a lookup needs one call; a "theology of" question needs the
full canon), discovery before thesis via survey,
mandatory attempts to falsify each thesis, and provenance
labels separating what is observed, what is counted, and what is
inferred — so answers come from the text, not from training data.
MCP server
$ claude mcp add bible -- bible mcp
bible mcp serves the same operations as typed MCP tools
over stdio for Claude Desktop, Cursor, and any MCP client.
Use it in the Claude app or ChatGPT — no install
https://mcp.biblecli.org/mcpA hosted endpoint serves the full toolset over Streamable HTTP. In the Claude apps (web and mobile): Settings → Connectors → Add custom connector, paste the URL. In ChatGPT: Settings → Connectors → Advanced → Developer mode, then add the URL as an MCP connector. The study methodology travels with the connection — the server delivers it to the model automatically, so any chat becomes a disciplined, text-first study session.
Command reference
| Reading | |
|---|---|
| passage <ref> | Read a passage. |
| read <scope> | Read the text sequentially, paged into context-sized chunks — study by reading, not just querying. |
| outline <book> | The shape of a whole book in one call: every chapter with its opening words, size, and most distinctive vocabulary. |
| study [command] | Inductive reading sessions: a durable cursor over a scope plus a verse-anchored notebook — read, observe, record; search follows observation. Start with: bible study start Genesis |
| search <query> | Full-text search. |
| compare <ref> | Compare a verse across translations. |
| Original languages | |
| interlinear <ref> | Word-by-word original language with English. |
| original <ref> | Original-language text of a passage. |
| syntax | Clause search over the MACULA treebanks: who did what to whom. |
| lemma <query> | Occurrences of a lemma or Strong's number. |
| word <query> | Word study: lexicon entries + usage stats. |
| morph <ref> | Full parse of every word in a verse. |
| grep-morph | Search by grammatical form. |
| variants <ref> | Textual variants for a verse or short range, repackaged from the tagged data: per-edition Greek texts with edition-disputed words (NT), Masoretic Ketiv/Qere and LXX-stream readings (OT), and alternate Hebrew/Greek versification. Printed-edition-level evidence, not a manuscript apparatus. |
| Pattern analysis | |
| survey <query> | Corpus dossier for a topic — run this FIRST in any study. Accepts a Strong's number, original-language lemma, English word, or passage. |
| quotes <ref> | OT-in-NT parallels computed from the Greek (LXX vs NT), in confidence tiers: quotation (5+ word run), allusion (4-word run), echo (shared rare vocabulary). |
| parallels <ref> | Inner-biblical parallels within a testament, computed from original-language lemma runs (Kings↔Chronicles, Psalm doublets, Synoptics, Jude↔2 Peter), in confidence tiers: parallel (5+ lemma run), allusion (4), echo (3 rare lemmas). |
| xref <ref> | Ranked cross-references. |
| freq | Frequency distribution. |
| cooccur [ref] | Co-occurrence analysis. |
| similar <ref> | Passages sharing distinctive vocabulary with a passage (idf-weighted lemma overlap; lexical, not semantic). |
| name <query> | Who/what is this? Individualised persons and places. |
| pattern | Original-language formula search: find verses where a sequence of Strong's numbers and/or original-script lemmas occurs in order, with observed-vs-expected concentration by book. Original-language only — English words are not accepted (find Strong's numbers with 'bible word'). |
| Introspection | |
| books | List the 66 books with codes, chapter counts, and verse-id ranges |
| translations | List available translations |
| editions | List Greek NT editions available for --edition filters |
| morph-codes | Explain the morphology fields and their possible values |
| licenses | Data sources, licenses, and required attributions |
| ref <text> | Parse and normalize a reference. |
| schema [table] | Show CREATE TABLE statements for the scripture databases (core + attached study/lxx/user), plus notes on verse-id encoding and query conventions. |
| sql <query> | Run a read-only SQL query against the scripture databases (core, plus study/lxx/user when installed, attached under those schema names). Discover tables with 'bible schema'. |
| Infrastructure | |
| db [action] | Manage the local databases: status | download | path |
| mcp | Run as an MCP (Model Context Protocol) server: stdio by default, or --http for remote connectors (Claude web/mobile, ChatGPT, ...) |
| agent-setup [harness] | Write the study-methodology guidance for an agent harness: claude | codex | opencode | generic |
| import <path> | Import a translation you have licensed access to (USFM folder/zip/file, or TSV lines of '<ref><TAB><text>') into a local-only database for personal study. |
| doctor | Diagnose the installation: platform, native SQLite driver, databases, integrity. Run this first when anything misbehaves. |
Every command documents itself: bible <command> --help.
Data & licensing
| Layer | Source | License |
|---|---|---|
| English translations | BSB (default), WEB, KJV, ASV | Public domain |
| Hebrew OT + morphology | STEPBible TAHOT (Leningrad text, ETCBC morphology) | CC BY 4.0 |
| Greek NT + morphology | STEPBible TAGNT (NA / TR / Byzantine / SBL marked) | CC BY 4.0 |
| Lexicons | Brown-Driver-Briggs (Enhanced), Abbott-Smith (ext.), Dodson | PD / CC BY / CC0 |
| Cross-references | OpenBible.info, ~345k vote-ranked | CC BY |
| Proper nouns | STEPBible TIPNR | CC BY 4.0 |
| Septuagint + quotation links (optional) | Swete edition + computed links | CC BY-SA 4.0 |
The databases are built by a reproducible pipeline with a
verification gate — canonical verse counts, versification edge cases,
morphology coverage — and full source attribution ships inside each
database: run bible licenses.