A production environment for scripture content: open biblical data and AI turned into published scholarly resources, with a documented trail from every claim back to its source.
Scripture Pipelines
The Problem
Biblical scholarship now has more open data than anyone can use — word-level annotation of the Hebrew Bible and Greek New Testament, treebanks, lexicons, discourse datasets, entity authorities.
The bottleneck is not data.
Very few people know how to use these resources together to create new ones. So the work that does happen is built from scratch each time: the same wheels reinvented, in bespoke code that is hard to maintain and harder to hand on.
Â
And general-purpose AI fails here in one specific way. It produces scholarship that sounds grounded in the text but was invented from training data. That failure is quiet, it is plausible, and it is disqualifying — no organization can publish what it cannot vouch for.
What It Does
Underneath the production framing it is a language. A pipeline states what should be produced and from what, so the person who commissioned the work can read it and check that it says what they asked for — which is not true of code. The language knows what the annotated texts, the lexicons, the entity authority and the discourse datasets mean, so an author works in terms of those resources rather than files and parsing.
Â
Every AI step must declare what data it requires before it runs, return structured output against a defined schema, and leave a record of exactly what it received and returned. Sources can be shown, and invented scholarship can be caught. Sign-off stays human: the environment verifies compliance, a scholar judges whether the output is fit to use.
Â
The catalog of sources is published as a public registry, Awesome Biblical Data — currently seventy datasets, useful to the field whether or not anyone runs a pipeline.
Where It Sits
Second in the stack. It is where the framework above it stops being intention and becomes machinery, and it produces everything below it. Open source, shipped for macOS, Windows and Linux, and in daily use at alpha maturity.
