The Fallible · Synthetic · Study Bible
A study Bible that shows its work — every claim sourced, every machine word marked, every gap stated.
The real artifact here is a provenance system — a discipline for separating what God said, from what a faithful human teacher said about it, from what a machine assembled. The Scriptures are the demonstration corpus precisely because they are the text where getting provenance right matters most. The same machinery could carry any body of contested knowledge. We point it at the Word because here, more than anywhere, a reader deserves to know exactly whose voice is speaking.
So this is not an AI that tells you what the Bible means. It is a frame that hands you the Word plainly, gathers the public-domain teachers around it by name, and adds its own machine reading last — clearly labeled, easy to ignore, never confused with the text.
Everything on every page belongs to exactly one of three tiers, and the tier is always visible. They are not equal, and we never let the design pretend they are.
The Berean Standard Bible (CC0, public domain), printed exactly. This alone has authority. Everything else on the page is commentary about it and may be wrong.
Commentary from named, public-domain teachers — quoted verbatim, attributed, dated, and linked to source. Their words are theirs; we do not paraphrase them into our own mouth.
This tool’s own work — literal renderings, word notes, the grand commentary, the canonical threads. Fallible and marked as such. Weigh it against the text; it has no authority at all.
The order is deliberate: the Word first, the saints of the church beside it, the machine last and lowest. A reader who ignores everything but the BSB has lost nothing essential.
Two glyphs run through the whole site. Learn them once and you can read any page at a glance:
Oxblood (✦) always means a human hand you can trace. Slate (⚙) always means a machine you must check.
The hardest and most original part is the Canonical Threads — the links out from a passage to the rest of Scripture. Most study Bibles simply assert these. We compute a basis for each one and show our reasoning, so you can disagree on the evidence rather than on our say-so.
Strong’s number — the lemma of the underlying Hebrew or Greek, independent of the English wording.basis:.When the basis is weak or impossible to compute, we say so out loud rather than bluff. The clearest case is a cross-Testament link: a Hebrew passage and a Greek one share no Strong’s number by definition, because they are different languages. So a claim that, say, Hebrews quotes Joshua cannot be proven by lexeme overlap. Rather than assert it, the Verifier flags it:
flagged — verify source
A flag is not a failure; it is the Verifier working in the open. It means: this connection is traditional and plausible, but the mechanical evidence is inconclusive, so go check it yourself. Contested links are left visible on purpose. We would rather show you a doubt than hide one.
The governing rule of the whole project: we never invent to fill a hole. Where the data stops, the page says so.
When the critical text omits a verse’s original-language words (a genuine textual variant), the page prints the BSB and shows no word cards — with a note that this is a stated gap, not invented text. When a voice hasn’t been gathered yet, the apparatus says the voices come in a later pass rather than fabricating a quotation. When a cross-reference can’t be mechanically verified, it is flagged, not asserted.
A reader should always be able to tell the difference between we checked and here it is, we checked and it’s uncertain, and we don’t have this yet. Blurring those three is the one thing this project refuses to do.
“They received the word with all readiness of mind, and searched the Scriptures daily, whether those things were so.”Acts 17:11