# CrossReferences.org

> An open, freely licensed corpus of phrase-level Bible cross-references, derived from the Treasury of Scripture Knowledge (TSK) and re-anchored to modern translations. Each reference is attached to a specific phrase in the target translation's own wording, and multi-verse references are kept whole rather than flattened into separate verse links.

Data first. The web Bible reader and the sandbox are demonstrations of what the corpus supports, not the product. Cross-reference data is CC BY 4.0, commercial use included.

## Who it's for

Anyone building or improving Bible software who needs cross-reference data: Bible reading and study apps, sermon preparation tools, offline and low-bandwidth apps, digital Bible publishing, and research on the text's internal connections. Useful in particular where an app displays a modern translation and wants to highlight the phrase a reference belongs to, or serves French or Afrikaans readers, or must work without a network connection. No API key, no account, no usage limits — download the files and embed them.

## What makes it different

Most free TSK derivatives drop the anchors and flatten the reference groups, leaving a verse-to-verse edge list. The TSK's precision was in its anchors: a reference attaches to a phrase within a verse, not to the whole verse. This dataset keeps that, rebuilt in each translation's own words.

Genesis 1:1, English (BSB) and French (S21):

```
Gen	1	1	beginning	Prov 8:22-24|Prov 16:4|Mark 13:19|John 1:1-3|Heb 1:10|1 John 1:1
Gn	1	1	commencement	Pr 8:22-24|Pr 16:4|Mc 13:19|Jn 1:1-3|Hé 1:10|1 Jn 1:1
```

There is an anchor column; it is in the target translation's wording, so an app showing a modern translation can highlight it directly; and `Prov 8:22-24` is one reference, not three.

63,668 anchors. About a third of reference groups span more than one verse. No dangling references. The set is also modestly curated — references added, split, merged or removed where it serves a reader, currently well under 1%.

## Translations

| Code | Translation | Language | Text licence |
|---|---|---|---|
| KJV | King James Version | English | Public domain |
| BSB | Berean Standard Bible | English | Public domain |
| S21 | Segond 21 | French | © Société Biblique de Genève |
| AOV | Afrikaanse Ou Vertaling | Afrikaans | Public domain |

**Segond 21:** this dataset does **not** contain the Segond 21 Bible text, and contains no S21 chapter or verse numbering. It carries S21 anchor phrases only — short fragments used to locate a reference within a verse. This is a deliberate licensing boundary. A French reader built on this data needs the S21 text from its rights holder, Société Biblique de Genève.

KJV, BSB and AOV texts are included and are public domain.

## Formats

**TSV** — one file per translation: `book`, `chapter`, `verse`, `anchor`, `references` (separated by `|`).

**JSON** — `bible_books.json`, `bible_verses.json` (the text layer), `cross_references.json` (anchors and reference groups, joined by verse id). `refs` is a list of lists: each inner list is one reference. Versification differences between translations are handled structurally; see the repository for the model and the QA scripts.

## Data

- [GitHub repository](https://github.com/CrossReferences-org/bible-cross-references) — source of truth, format documentation, QA scripts
- [Latest release](https://github.com/CrossReferences-org/bible-cross-references/releases/latest) — versioned snapshot, with release notes
- [Rolling latest](https://github.com/CrossReferences-org/bible-cross-references/archive/refs/heads/main.zip) — current state of `main`
- [Licensing](https://crossreferences.org/project/licensing/)

Data files are generated exports, so pull requests against them cannot be merged. Corrections are welcome as issues.

## Tools

- [Bible reader](https://crossreferences.org/tsk/) — select a phrase, see the passages it connects to with full text. No account, no ads, works on slow connections
- [Connection Explorer](https://crossreferences.org/sandbox/connection-explorer) — experiment that walks the reference graph several hops out

## What this is not

- **Not an API.** No hosted endpoint for third-party apps, and none planned. The data is downloaded and embedded, so an app keeps working independently of this site.
- **Not paywalled.** No account, no fee, no gate.
- **Not a source for the Segond 21 text.** Anchor phrases only.

Cross-reference data © CrossReferences.org, [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). Bible texts remain under their own licences. The licence covers the dataset and places no requirement on code that uses it.