An AI-free core
The retrieval and attribution core runs with no language model in the path. pericopa references Scripture; it never produces it. No paraphrase, no synthesis, no generated text standing between you and the source.
Reference, not generation
pericopa is the trust layer for Bible study: it surfaces where a passage lives, which tradition's canon holds it, and where editions diverge — every reference traceable to a named source.
The thesis
Most tools answer a Bible question by handing you a sentence and asking you to trust the tool. pericopa does the opposite: it hands you the citation. Each reference carries its source, its tradition, and its points of divergence, so you can follow it back to the record yourself instead of taking a single rendering on faith.
The retrieval and attribution core runs with no language model in the path. pericopa references Scripture; it never produces it. No paraphrase, no synthesis, no generated text standing between you and the source.
Canon, versification, and tradition are structural, not footnotes. pericopa is built on a schema-validated doctrine model, so a reference always carries the tradition it belongs to — never flattened into one default Bible.
A citation is only as good as the path back to its source. pericopa keeps that path intact end to end, so any reference it surfaces can be followed to the named edition it came from.
The apparatus
pericopa borrows the discipline of a critical apparatus. Its marks name what you are looking at; they do not grade it. Color is an identifier here, never a judgement.
Marks where editions of a passage diverge — wording, versification, what each one includes. A reading lens over the textual record, not a position on it.
edition diverges here