Guides

Greek and Hebrew data for programmers

4 min read · last updated 2026-11-18

There is a surprising amount of free, high-quality data on the Greek and Hebrew text of the Bible. Every word tagged with its dictionary form and grammar. Syntax trees for whole books. Lexicons. A lot of it used to be locked up in expensive software, and over the last several years much of it has been released under open licenses.

If you want to build a word study tool, an interlinear, a flashcard app for Greek vocabulary, or just play with the data, this is where to start.

Greek New Testament

The base text most open projects use is the SBL Greek New Testament (SBLGNT). It’s a critical edition from the Society of Biblical Literature, and it’s licensed under Creative Commons Attribution 4.0 (CC BY 4.0). That means you can use it for almost anything, as long as you give credit.

MorphGNT adds grammar tags to the SBLGNT. It’s about as easy to parse as data gets. Each word is one line, with columns separated by spaces:

010101 N- ----NSF- Βίβλος Βίβλος βίβλος βίβλος
010101 N- ----GSF- γενέσεως γενέσεως γενέσεως γένεσις

The first column is book, chapter and verse (book 01 here is Matthew, the first book of the New Testament). Then part of speech (N- is a noun), the parsing code (nominative singular feminine, then genitive singular feminine), the word as it appears in the text, the word without punctuation, a normalized form, and finally the lemma, the dictionary form you’d look up. You can load the whole New Testament into memory with a few lines of code.

MACULA Greek goes a lot further. It has full syntax trees, so you know which words are the subject of which verb, plus English glosses, word senses and more. It covers both the SBLGNT and the Nestle 1904 text, comes in XML and TSV, and is licensed CC BY 4.0. Some of the sources bundled with it have their own terms, so read its license file before you redistribute everything.

Hebrew Bible

For the Old Testament, the base text is usually the Westminster Leningrad Codex (WLC), which is in the public domain.

The Open Scriptures Hebrew Bible (OSHB) tags every word of the WLC with a lemma and morphology code. It’s in OSIS XML, and a word looks like this:

<w lemma="b/7225" morph="HR/Ncfsa">בְּ/רֵאשִׁ֖ית</w>

That’s the first word of Genesis, “in the beginning.” The slash splits the prefix (the preposition b, “in”) from the main word. The lemma number is based on Strong’s numbering, and the morph code says it’s Hebrew, a preposition, then a common noun, feminine singular absolute. The morphology and lemma data is CC BY 4.0.

MACULA Hebrew does for the Hebrew Bible what MACULA Greek does for the New Testament: syntax trees and linguistic annotations, CC BY 4.0, built on the WLC and OSHB.

STEPBible data

Tyndale House in Cambridge releases the data behind STEPBible.org on GitHub under CC BY 4.0. It includes:

  1. Tagged Greek and Hebrew texts that also note where manuscripts and editions differ.

  2. Lexicons for Greek and Hebrew based on an extended version of Strong’s numbers, which fixes a lot of the places where the original Strong’s numbering lumps different words together.

  3. Proper noun data, versification mappings, and morphology code explanations.

Most of it is in tab-separated text files. They’re big and have a lot of columns, so read the README in each folder first. If you want one source that ties texts, lexicons and translations together, this is probably it.

Which one should you start with?

If you just want to count words or build a vocab list, start with MorphGNT. You’ll be working with real data in ten minutes.

If you need lexicon definitions, use the STEPBible lexicons.

If you need to know how words relate to each other in a sentence, use MACULA.

For the Old Testament, start with OSHB, then add MACULA Hebrew or the STEPBible data as you need them.

A word about Strong’s numbers

You’ll see Strong’s numbers everywhere. They come from James Strong’s concordance from the 1890s, and they’re still the most common way to link words across tools. They’re also kind of rough. Some numbers cover several different words, and some words got split across numbers. The extended numbering in OSHB and STEPBible fixes many of those problems, so if you have a choice, use the extended IDs and keep the plain Strong’s number as a secondary field.

Give credit

Almost all of this data is CC BY, which only asks one thing: give credit. Put an “about” or “data sources” page in your app that names each dataset, links to it and states its license. It’s the right thing to do, and it’s what the license requires.

If you build something with this data, show it off in Show & Tell. People who care about this stuff are usually really excited to see it used.


Questions or corrections? Ask in the forum.

Other guides