About Diviane
An open-source platform for exploring the Bible — hybrid vector and graph retrieval, with every generated claim grounded in a cited source.
About the name
Diviane reads, in French, as div(IA)ne — "IA" (French for AI) hidden inside "divine." It's also, independently, a real name in the Diane / Viviane family, which is exactly why it works in English too, even though the AI pun itself is French-specific and won't be obvious to English-only readers. Both readings are true; neither needs the other to work.
Editorial principles
- No modern copyrighted translation is ever ingested, quoted, or reproduced. WEB and other public-domain/openly-licensed texts only.
- No single doctrinal position is presented as "the" Diviane position. Interpretation cards are sourced and multi-perspective by construction.
- Traditional chronologies (e.g., Ussher) are labeled as historical/traditional artifacts, not as scientific consensus or fact, consistent with how they are understood in contemporary scholarship.
- The introspection/journaling feature is reflective, not clinical. It does not diagnose, does not simulate therapy or pastoral counsel, and will point users toward real human support for anything beyond reflective journaling.
- What is stored, and where, is written down. Journal entries never leave your browser — no accounts, no server-side journals, no analytics — and the transient caches that do exist expire within 24 hours; the full inventory is
docs/PRIVACY.md. - Every generated claim is grounded and cited. If retrieval doesn't surface a source for a claim, the claim is not generated.
- Rewrites are always paired with, never a replacement for, the source text.
Credits & licenses
Diviane’s own code is AGPL-3.0. The Bible text is public domain. The structured datasets are openly licensed, and the credit lines their licenses require are below. The full per-source record, with licenses, pins and fetch dates, is the project’s data ledger, data/SOURCES.md. Interpretation cards carry their own source’s citation beside every quote.
BibleData
Person, Place, Event, Epoch, PersonRelationship, PersonVerse, the 613 commandments, HebrewStrongs, Hitchcock's Bible Names Dictionary and the Alamo Polyglot — the knowledge graph's structured backbone.
Brady Stephenson. (2026). BibleData: Structured Datasets from the Holy Bible (Version 1.0) [Data set]. Zenodo. https://doi.org/10.5281/zenodo.19539956
CC BY 4.0 Source
Per-file contributors credited within BibleData: Dan Raby (PersonLabel) and Fernando Falci (PersonRelationship).
Ussher's Annals of the World
The traditional 17th-century chronology, served as a labeled traditional artifact — never as scientific consensus or fact.
Brady Stephenson. (2026). BibleData: Structured Datasets from the Holy Bible (Version 1.0) [Data set]. Zenodo. https://doi.org/10.5281/zenodo.19539956
CC BY 4.0 — the same terms as BibleData, which is where it ships from Source
OpenBible.info cross-references
The ~345k verse cross-references behind a verse's cross-reference list.
Cross references from OpenBible.info (www.openbible.info), used under CC BY 4.0.
CC BY 4.0 Source
OpenBible.info Bible Geocoding
The place identifications and coordinates behind place pages.
Bible Geocoding data from OpenBible.info (https://www.openbible.info/geo/), used under CC BY 4.0.
CC BY 4.0 Source
Open Scriptures Hebrew Bible (OSHB)
The Hebrew Masoretic text's lemmatization and morphology, read by the interlinear.
Original work of the Open Scriptures Hebrew Bible available at https://github.com/openscriptures/morphhb
CC BY 4.0 — for the annotations; the Westminster Leningrad Codex base text is public domain Source
SBL Greek New Testament (SBLGNT)
The Greek New Testament critical text the interlinear reads.
The SBLGNT is licensed under a Creative Commons Attribution 4.0 International License. Copyright 2010 by the Society of Biblical Literature and Logos Bible Software.
CC BY 4.0 Source
STEPBible TAGNT
Word-level tagging of the Greek New Testament: extended Strong's numbers, morphology, transliteration and glosses.
Data created by www.STEPBible.org based on work at Tyndale House Cambridge (CC BY 4.0).
CC BY 4.0 Source
STEPBible TIPNR
The proper-noun layer: people and places with their disambiguated names and the references to them.
Data created by www.STEPBible.org based on work at Tyndale House Cambridge (CC BY 4.0).
CC BY 4.0 Source
World English Bible (WEB)
Every verse of scripture the reader shows.
World English Bible text (public domain, ebible.org), in the structured JSON packaging by TehShrike.
Public domain — no credit required, credited anyway Source
Strong's Greek Dictionary
The Greek lexicon entries behind a Strong's lookup and the interlinear's per-word glosses.
Strong's Greek Dictionary in XML with real Greek, prepared by Ulrik Sandborg-Petersen.
CC0 1.0 — no credit required, credited anyway Source