TrulyRandomVerse MCP Server

Fetch the full lexical entry and word study for a single Strong's number (e.g. 'H7563', 'G2316'). Returns lemma, transliteration, pronunciation, KJV definition, `word_study`, and `grammatical_forms` — the word's part of speech plus how it actually inflects across all its occurrences in Scripture (each form with its occurrence count, most frequent first). Use `grammatical_forms` to ground claims about how a word is used grammatically (e.g. "most often a Piel — an intensive stem") rather than generalizing from memory. `word_study` is an essay written with AI in 2025 and partly reviewed since: the definition, forms and occurrence count are the recorded data, and a verse the essay cites for a word may not contain it, so check with get_strongs_for_verse or get_strongs_occurrences before saying it does. Call this AFTER get_strongs_for_verse when you need to go deeper on a specific word.

List every Strong's number tagged in a verse along with the original Hebrew/Greek lemma, transliteration, short gloss, and — for inflected words — its exact grammatical parse IN THIS VERSE (`grammar` field: e.g. "Present Passive Deponent Participle Nominative Plural Masculine" for a verb, "Accusative Singular Feminine" for a noun). Use this when the user asks about original language, word meaning, or grammar, or when you want to do word-level analysis. ALWAYS ground any claim about a word's tense/voice/mood/case/number/gender in this `grammar` field — never parse the Hebrew/Greek yourself from memory. Words with no `grammar` field are indeclinable (particles, conjunctions) and have no parse. The result also carries `interlinear`: the verse in its ORIGINAL word order — each token's surface form exactly as written (not the lemma), transliteration, contextual English gloss, Strong's link, and a compact morphology code (Hebrew e.g. "V-Qal-Perf-3cp", Greek Robinson e.g. "V-AAI-3S") plus `parse`, that code spelled out for THIS token specifically; compound words carry `segments` splitting prefixes from the content word, each flagged `is_content` — true marks the root, false a prefix or suffix. Ground any claim about word ORDER, emphasis-by-position, repetition, or what the original "literally says" in `interlinear.words`, quoting surface forms from it — never reconstruct the original text from memory. In Psalms, `interlinear.superscription_words` = how many leading words are the psalm's superscription/title (the KJV folds it into verse 1).

List every verse where one Hebrew or Greek word occurs, by its Strong's number: a concordance. Each verse comes with its text and, where the interlinear has it, `original`: the word as written in that verse, its transliteration, its gloss there and its grammatical parse there. Use it to see how Scripture itself uses a word, and to find passages that explain each other by sharing one (2 Samuel 24:1 and 1 Chronicles 21:1 both use H5496, "moved" and "provoked"). `total` counts VERSES (in `book`, when given); `uses_in_scripture` counts every use of the word in the whole Bible, since one verse can hold it twice (each verse's `original` lists every use there). Up to 40 verses come back per call, in canonical order. When there is more than one page, `by_book` shows where all the verses fall; call again with `next_offset`, or pass `book` to read one book. Pass `followed_by` to find a phrase in the original, whatever the English: H1121 followed by H430 is bᵉnê hāʼĕlōhîm, "sons of God" (Genesis 6:2, Job 1:6); add `gap` to allow words between. Verse text is in the user's conversation language by default; `lang` overrides. Outside trulyrandomverse.com/ai the default edition is the KJV ('english'); pass `lang` for another. Editions whose licence does not allow redistribution are refused. Book, chapter and verse are always KJV numbering, in every edition: the text returned is that edition's rendering of the same KJV verse. A few editions number some chapters differently in print (Ukrainian, Douay-Rheims, Geneva, Swedish and Serbian, mostly in Psalms), so a reader quoting their own Bible's number there may mean a neighbouring KJV verse.

A whole book of the Bible at a glance: who wrote it and when (where faithful Christians differ, each view with its grounds, given as references and the Church Fathers' own words), how the book is laid out in movements, its key verse, the Hebrew or Greek words it turns on (Strong's numbers), its themes, how it points to Christ, its links to other books, what the Fathers said of it, and its main people and places. CALL THIS FIRST for a question about a whole book — who wrote it, when, to whom, how it is built, what it is about, how to read it — then go deeper with the verse and word tools. It covers the 66 books of the Protestant canon. The result's `url` is the book's study page on trulyrandomverse.com; to send the reader there, link that address.

Fetch quotes from the early church fathers (Augustine, Chrysostom, Origen, Basil, Ephrem, etc.) on a verse or verse range. Returns each quote verbatim: whole up to 4,000 characters; a longer one as an excerpt of the parts that bear most on the verse, marked `excerpt` (search_fathers with its `quote_id` returns it whole). The same writer's identical text filed under two verses of the range comes once, the other in `also_under`. Pass `verse` alone for quotes on that verse. Pass `verse_end` ALSO to sweep a passage; a sweep lists quotes in verse order, each `reference` naming the verse or verses that quote comments on. When more quotes exist than `limit`, each writer gives one before any writer gives a second, the centuries taking turns, so the answer carries as many voices and eras as it can. `total` says how many exist; when that is more than `limit`, `quotes_by_verse` shows where they all fall — never say a verse has no patristic comment until `quotes_by_verse` or a call on that verse alone confirms it. If the requested range produces >200K chars of quotes total, the tool returns `mode: range_too_large` with a focus prompt — ASK THE USER which passage to focus on rather than calling repeatedly. Returns father name, year AD, and the quoted text; a quote whose writer is not a Christian commentator (Philo and Josephus, Jewish writers; a book of the Bible filed as a cross-reference) carries `writer_note` — never call them church fathers. Quote directly with attribution; never paraphrase from training. Call it when the question concerns a passage's meaning, theology, or historical interpretation.

Fetch a section of the curated in-house commentary for a verse. Sections: 'summary', 'context', 'exposition', 'theological', 'reflection', 'faq', 'christ_centered'. One of three peer commentary sources — pair with get_matthew_henry and get_church_fathers for a full picture. None is sufficient alone; do not treat this as 'the' source.

Compare a verse (or short passage) across multiple Bible translations side-by-side. Use this when the user asks how one version renders something differently from another — e.g. 'how does the ASV/BSB say…', 'compare KJV and YLT on…', or when a translation choice itself is the topic. Best for English where multiple translations are available; non-English versions can be included too. Use 2-5 versions per call; do NOT call this just to fetch a single version (use get_verse_text for that). If the user asks for ALL or EVERY English version (or any superset of >5 versions), you MUST issue this tool MULTIPLE times in the SAME batch — e.g. one call with versions 1-5 and a second parallel call with versions 6-8 — until every requested version is covered. Never write about a translation that was not returned by a tool result this turn. Outside trulyrandomverse.com/ai the default edition is the KJV ('english'); pass `lang` for another. Editions whose licence does not allow redistribution are refused. Book, chapter and verse are always KJV numbering, in every edition: the text returned is that edition's rendering of the same KJV verse. A few editions number some chapters differently in print (Ukrainian, Douay-Rheims, Geneva, Swedish and Serbian, mostly in Psalms), so a reader quoting their own Bible's number there may mean a neighbouring KJV verse.

Get cross-references for a verse: OpenBible.info's, the same list the verse page shows, ranked by its readers' votes (each carries `votes`), counting links listed under either verse. Returns the referenced verses with their text included so you can cite scripture-with-scripture without a second lookup; `total` says how many exist. These link passages by theme and teaching; for passages that share an original Hebrew or Greek word, use get_strongs_occurrences (with `followed_by` for a two-word phrase). Verse text comes back in the user's conversation language by default; pass lang to override. Each cross-ref may carry a `note: "fallback: english"` if the localized translation is missing. Outside trulyrandomverse.com/ai the default edition is the KJV ('english'); pass `lang` for another. Editions whose licence does not allow redistribution are refused. Book, chapter and verse are always KJV numbering, in every edition: the text returned is that edition's rendering of the same KJV verse. A few editions number some chapters differently in print (Ukrainian, Douay-Rheims, Geneva, Swedish and Serbian, mostly in Psalms), so a reader quoting their own Bible's number there may mean a neighbouring KJV verse.

Search the patristic corpus BY WRITER rather than by verse — the corpus is 62,484 passages from 314 writers. Use this for questions ABOUT a father instead of about a verse: 'what did Chrysostom write on?', 'which books of the Bible did Augustine comment on?', 'what did Jerome say about fasting?', 'who commented on Habakkuk?'. Pass `father` alone for a profile plus the full list of biblical books they commented on, with counts. Pass `father` AND `topic` to search inside that writer's work. Pass `topic` alone to search all 314 writers at once. Pass `father` AND `book` for their commentary on one book. Pass nothing for the list of writers the corpus carries. Quotes returned are verbatim from the curated DB — quote them directly with attribution and never paraphrase from training. When the writer's name is ambiguous (there are four Gregorys) the tool says so and lists the candidates; ask the user which one rather than picking. Each quote carries its `quote_id`; a long one comes as an excerpt (`excerpt` names its parts), and `quote_id` alone returns a quote whole, from any tool that gave the id. The same writer's identical text filed under several verses comes once, the others in `also_under`.

Fast keyword search of the KJV (English index — phrase your query in English). Two or more words are read as a phrase first: verses with those words in that order come first, in Bible order, with `phrase_total` counting them all; then verses with every word; then related verses ranked by verse text and the site's commentary. Each verse's `match` says which (phrase, all_words, related). Returned verse text is in the user's conversation language when available, KJV otherwise. Outside trulyrandomverse.com/ai the default edition is the KJV ('english'); pass `lang` for another. Editions whose licence does not allow redistribution are refused. Book, chapter and verse are always KJV numbering, in every edition: the text returned is that edition's rendering of the same KJV verse. A few editions number some chapters differently in print (Ukrainian, Douay-Rheims, Geneva, Swedish and Serbian, mostly in Psalms), so a reader quoting their own Bible's number there may mean a neighbouring KJV verse.

Fetch Matthew Henry's commentary on a verse or verse range — a foundational, widely-cited Protestant exegetical work. Returns the full text MH wrote, never truncated. Pass `verse` alone for the chunk covering that verse (MH commentates on ranges, so this often returns content for surrounding verses too). Pass `verse_end` ALSO for a chapter or passage sweep — every overlapping chunk at full text. If the requested range is too large (>200K chars total, e.g. all of Psalm 119, or a whole-book request), the tool returns `mode: range_too_large` with a focus prompt — ASK THE USER which chapter or passage to focus on rather than calling repeatedly. Quote directly with attribution; never paraphrase from training. Peer of get_commentary_section, not a fallback. Call it when the question concerns a verse's meaning, theology, or application.

Who a biblical name actually belongs to. Returns the distinct individuals and places bearing a name — each with its own identity, description, tribe or nation, family (parents, siblings, partners, offspring, every one carrying the related person's identity so you can look them up in turn), and the verses they appear in. Places carry coordinates. CALL THIS WHENEVER A QUESTION IS ABOUT A PERSON OR PLACE RATHER THAN A WORD: who someone was, whose son or daughter, who they married, where they appear, whether two mentions are the same man. CRITICAL: many biblical names are shared by several unrelated people — one name in this data covers twenty-nine different men — so this tool returns ALL of them and you must not assume the famous one is meant. If the question concerns a particular verse, match that verse against each individual's references before answering, and if it stays ambiguous, say so. Uncertainty in the source is preserved deliberately: a '(?)' marks a relationship the data itself is unsure of, and 'Unnamed#N' is a person Scripture never names — report both as they stand rather than smoothing them over. For what a WORD means or where it came from, use get_word_origin instead.

Answer questions about THIS WEBSITE (TrulyRandomVerse / trulyrandomverse.com) and its apps — what a feature is, where to find it, how to use it, why it behaved a certain way, or what the site's policy is. Covers the browser extension, installing as an app, accounts and syncing, favorites, the Ten Lists tracker, search, privacy, donations, and every reading and study page. Call it whenever the user asks about the site itself rather than about Scripture ('how do I…', 'where is…', 'does this site…', 'why did my… change', 'what is this site', 'how do I install the extension'). You know NOTHING about this site's features except what this tool returns — never answer a site question from memory or by guessing what a Bible site probably does, and never assume a feature exists because it would be sensible. Returns curated help text plus the page to link. To send the reader to that page, link its `url`. Request up to 3 topics in one call.

Find verses tagged under a biblical topic. Use this when the user asks about a theme (e.g. "covenant", "judgment", "grace") and you want canonical examples from across Scripture. Verse text comes back in the user's conversation language by default; pass lang to override. Each verse may carry a `note: "fallback: english"` if the localized translation is missing. Outside trulyrandomverse.com/ai the default edition is the KJV ('english'); pass `lang` for another. Editions whose licence does not allow redistribution are refused. Book, chapter and verse are always KJV numbering, in every edition: the text returned is that edition's rendering of the same KJV verse. A few editions number some chapters differently in print (Ukrainian, Douay-Rheims, Geneva, Swedish and Serbian, mostly in Psalms), so a reader quoting their own Bible's number there may mean a neighbouring KJV verse.

Fetch the raw verse text of a single verse or a small range (max 10 verses). By default returns text in the user's conversation language; pass lang='english' (or another supported code) to override. If the requested language has no translation row for a verse, the result includes a `note: "fallback: english"` field for that verse and returns the KJV text — translate it as you cite. Outside trulyrandomverse.com/ai the default edition is the KJV ('english'); pass `lang` for another. Editions whose licence does not allow redistribution are refused. Book, chapter and verse are always KJV numbering, in every edition: the text returned is that edition's rendering of the same KJV verse. A few editions number some chapters differently in print (Ukrainian, Douay-Rheims, Geneva, Swedish and Serbian, mostly in Psalms), so a reader quoting their own Bible's number there may mean a neighbouring KJV verse.

Trace where a Hebrew or Greek word came from and what the language built out of it. Returns `formed_from` (the word or words this one is built out of — each with its own lemma, transliteration and definition, so a compound can be read part by part), `words_formed_from_this` (the reverse direction), `related_language_entry` (the matching entry in the other testament's language, where the lexicon records one — this is how an Old Testament name and its New Testament spelling are connected), whether Scripture uses the word as a proper name, and where it occurs. Call this when someone asks what a word or name MEANS at its root, where it comes from, whether two words are related, or how a name reached its familiar English spelling. IMPORTANT: everything returned here is recorded lexical data the reader can check. Comparative or historical claims that do NOT appear in these fields — cognates in other ancient languages, reconstructed meanings, scholarly proposals — are outside this data: you may still discuss them, but say plainly that they are scholarly argument rather than something the lexicon records.