Isaiah 7:14 is probably the most argued verse in OT textual criticism. MT has עַלְמָה, LXX has παρθένος, and the 1QIsaᵃ scroll at Qumran agrees with MT. Patristic writers from Justin Martyr through Jerome built entire Christological arguments on the Greek reading. The question of whether that choice was translation technique, theological tendency, or something else has filled monographs.
I wanted to see what a tool grounded in the actual corpora, not just Claude's training knowledge would do with it. So I built one.
BibCrit (bibcrit.app) is a free, browser-based textual criticism toolkit. No login. Open-source (Apache 2.0). It runs analysis against real corpus data: ETCBC morphological database for the MT, STEP Bible for the LXX, ETCBC DSS modules for 1QIsaᵃ and three other scrolls, Samaritan Pentateuch, Peshitta, Targums (via Sefaria), and the Clementine Vulgate. The AI analysis is explicitly prompted to apply the frameworks of Tov, Ulrich, and Metzger — not just answer generically.
Fourteen tools at the moment:
- MT/LXX Divergence Analyzer — word-level classification (different Vorlage, theological tendency, scribal error, translation technique) with competing hypotheses ranked by confidence
- Hebrew Vorlage Reconstruction — retroversion of LXX readings back to probable Hebrew with per-word confidence scores
- Ancient Witness Bridge — aligns DSS, SP, Peshitta, MT, and LXX for a given passage
- Scribal Tendency Profiler — five-axis analysis per LXX book (literalness, anthropomorphism reduction, messianic heightening, harmonization, paraphrase rate)
- Source Criticism — J/E/D/P attribution via divine name usage, doublets, vocabulary patterns
- Patristic Citation Tracker — Church Father text-form distribution through the 5th century
- Plus chiasm detection, numerical discrepancy modelling, manuscript genealogy, Targum comparator, NT textual tradition, Second Temple Literature intertextual mapping, and a few others
Results stream in real time. Everything is exportable as SBL footnotes or BibTeX. There's an open cache API if you want to pull results programmatically.
The tool is grounded enough that I think it's useful for actual research — at minimum as a fast first-pass before you go to Tov's apparatus or the Göttingen edition. But I'd rather hear from people who know this material better than I do. Happy to answer questions about methodology, corpus sourcing, or how any specific tool works.