orthotomeo

“Cutting straight the word of truth.”

- 2 Timothy 2:15

A scripture-study engine, cited down to the row.

Look up any word in the Greek New Testament, Hebrew Old Testament, or Septuagint and see every place it occurs - the original text, its grammar, and the manuscripts behind it, each traced to the exact line in a real source file.

Use it here, or connect it to an AI assistant over MCP - so what it tells you about the text is retrieved from the corpus, not recalled from training. Nothing summarized, nothing paraphrased, nothing invented.

Ways to study

Interlinear

Read a verse in its original language, lined up word for word with an English translation - lemma, gloss, and grammar for every word.

Morphology

See exactly how a Greek or Hebrew word is formed - tense, case, mood, stem - so you can tell what the grammar is actually doing.

Concordance

Find every place a word or lemma occurs across the whole corpus. It returns all of them or it fails - never a quiet, partial list.

Lexicon

Look up a word's dictionary entry and glosses, each cited to the edition it came from.

What you can rely on

Traceable

Every citation points back to a real line in a real source file. You can always check the receipt.

Complete, or it says so

A search returns every match, or it tells you it couldn't - it never trims the list to look tidy.

It shows, never argues

The engine hands you the words and where they occur. It never spins a position or tells you what to conclude - that stays your work.

Why it's built this way

Romans 3:4

“Let God be true, but every man a liar.”

When a dictionary, a grammar, a commentary, or an AI's confident guess disagrees with the text, the text wins. The engine keeps no shortcut for deciding what a word “really means” - it hands you every occurrence and lets scripture settle scripture.

1 Peter 4:11

“If any man speak, let him speak as the oracles of God.”

Anything said about scripture is answerable to scripture. So every citation here traces to an actual line in an actual source file - never a summary, a paraphrase, or something half-remembered.

2 Timothy 2:15

“Rightly dividing the word of truth.”

“Cutting straight” is a charge to handle the text with care and precision. That's why the core operation is complete-or-fail: every occurrence, or an honest error - never a best-effort guess.

A worked example

A real exchange with an AI assistant connected to orthotomeo. Every claim it makes traces to a tool result; where it reasons past the data, it says so out loud.

“Where does ἀρθοτομέω occur in the New Testament, and what does it mean in context?”

Exactly one occurrence: 2 Timothy 2:15 - ἀρθοτομοῦντα τὸν λόγον τῆς ἀληθείας. A New Testament hapax - it appears nowhere else, so there's no second NT instance to fix its sense against.
2TI.2.15 · V-PAP-ASM
attestation: NKO
NA28+NA27+Tyn+SBL+WH+Treg+TR+Byz

“Search the Septuagint too.”

Two occurrences, both in Proverbs - 3:6 and 11:5 - each rendering a form of the same verb. Swete's Septuagint is a surface text with no lemma tags, so a lemma search against it correctly returns zero, not a false blank.
PRO.3.6, PRO.11.5
lemma: ἀρθοτομέω · OSS-LXX-lemma
Swete (surface-only): 0 matches

“Now cross-reference the underlying Hebrew.”

Both Proverbs verses translate the same Hebrew root, יָשַׁר (yashar, Piel) - “to make straight or level.” The “cutting” image is the Greek translator's choice, not something in the Hebrew. Worth weighing when you read 2 Timothy 2:15.
יָשַׁר
H3474 · Piel · “make straight/level”
xref: TAHOT, PRO.3.6 / PRO.11.5

Three Greek instances, two underlying Hebrew ones, nothing else in the tagged corpus. The synthesis above is inference from the pattern - kept explicitly distinct, throughout the real exchange, from what any single citation states outright.

The prompt behind it

The instructions given to the assistant for the exchange above. Copy it as a starting point for your own study.

You have access to the orthotomeo tools over a real biblical-text corpus. When I ask a scripture question, answer using the concordance method: pull every relevant occurrence - not a remembered sample - and let the unambiguous cases fix the ambiguous ones. Work from what the tools return, not from recall; if a grammatical or lexical claim isn't something you can point to in the returned citations, say so rather than asserting it. This includes when you derive a form yourself (for example, lemmatizing a surface-only edition that carries no lemma field): that's fine to do, but say plainly that you did it, distinct from citations the tool verified directly. Stay strictly with the biblical text itself - no commentaries, no modern theological arguments, no extra-biblical sources unless I ask for them as a separate, clearly marked follow-up. Report manuscript attestation as neutral data - a variant is mine to weigh, not yours to argue. Give me a reference list with brief lexical notes I can take into my own study, not a finished conclusion.

Use it inside Claude · MCP

What this is

MCP is the standard that lets an AI assistant use outside tools. Connect orthotomeo once and you can ask Claude a scripture question in plain language - it looks up the real text through the engine instead of leaning on memory, and every answer comes back cited.

  1. Open Claude on a plan that supports connectors (claude.ai or the desktop app).
  2. Go to Settings → Connectors → Add custom connector.
  3. Paste the URL on the right and save.
  4. Ask away - “find every place ἀγάπη occurs in the New Testament.”

Some AI assistants require a paid subscription to connect to MCP. No account, sign-in, cookies or tracking on orthotomeo's side - it's read-only and public.

Server URL
https://orthotomeo.app/mcp
Or, in a config file
{
  "mcpServers": {
    "orthotomeo": {
      "url": "https://orthotomeo.app/mcp"
    }
  }
}

Questions

Can I use this with other AI assistants, like ChatGPT?

Yes, if your assistant supports remote MCP servers - the connection is a URL, and it's the same URL for every client. orthotomeo implements the protocol; it doesn't care which assistant is on the other end. Which clients support remote MCP servers changes often enough that it's worth checking your own rather than trusting a list here. If yours doesn't, there's a plain GET JSON API that any tool can call.

Do I need an account or an API key?

No. There's no sign-in, no key, no cookies and no analytics - nothing about you is tracked, profiled or shared, and the corpus is public-domain and openly licensed text, so there's nothing to gate. The server keeps ordinary request logs, as any web server does, and the public instance is rate limited per day to keep it available.

What happens if my search matches thousands of verses?

It's refused, with the exact number it would have returned. It's never quietly trimmed to a sample, because a partial concordance that looks complete is worse than no answer - it's the failure the whole engine exists to prevent. Ask for a count first, then narrow.

Does it interpret the text, or take a position?

No. It returns words, grammar, manuscript attestation and where things occur. Textual variants are reported as data for you to weigh, not argued for or against. The reasoning stays yours; the engine's job is to make sure it rests on what the text actually says.

Why does the Septuagint sometimes not line up with the English?

Because in places it's a different text, not the same text renumbered. Greek Jeremiah is shorter than the Hebrew and its chapters run in a different order, so for that stretch orthotomeo asserts no verse-level correspondence at all rather than guessing one. You get an explicit note instead of a confident wrong verse.

Can I run it myself, or offline?

Yes. The engine is MIT licensed and the source is on GitHub, with command-line and desktop builds besides the web tool. The corpus editions keep their own licences, which are listed below.

Why isn't the ESV, NIV or NASB here?

Licensing. Those translations are under copyright and can't be redistributed as a database, which is what this engine is - every verse stored locally, queryable, and served with its source file and locator. An API that permits a few verses at a time can't support a complete-or-fail concordance. If you want a modern, readable translation, the Berean Standard Bible is here and is a good one: it was placed in the public domain in 2023 and is translated from the critical text.

The corpus

Every edition here is public domain or openly licensed, and that is a requirement rather than a preference: this engine stores the whole text locally and serves it back with its source file and locator, which a copyrighted translation's licence does not permit. It is why some familiar modern translations are absent, and why the Berean Standard Bible - public domain since 2023 - is the modern English edition on the list.

TAGNT - Translators Amalgamated Greek New TestamentCC BY 4.0 · STEPBible / Tyndale House
TAHOT - Translators Amalgamated Hebrew Old TestamentCC BY 4.0 · STEPBible / Tyndale House
TBESG / TBESH - Strong's lexicons (Greek / Hebrew)CC BY 4.0 · STEPBible
Swete's SeptuagintPublic Domain · archive.org
OSS-LXX-lemma - Open Scriptures LXX lemma indexCC BY 4.0 · Open Scriptures
BSB - Berean Standard BiblePublic Domain · Bible Hub
KJV / ASV / WEB / BrentonPublic Domain
OpenBible cross-referencesCC BY 4.0 · OpenBible.info