BibCrit (v.3.30) An AI-Assisted Web Application for Biblical Textual Criticism

What's new in 3.3.0 LXX Manuscript Witnesses A new analysis tool at /lxx-witnesses — the fifteenth BibCrit tool. Where the great Septuagint uncials diverge at a passage, and why it matters: Vaticanus (B), Sinaiticus (ℵ), and Alexandrinus (A). It uses Rahlfs-Hanhart corpus text where available plus Claude's knowledge of the critical apparatus. What it does Identifies significant inter-codex divergences per passage — locus, competing readings, and text-critical significance Each divergence and witness carries a confidence score Flags witnesses that are not extant (e.g. Sinaiticus' Octateuch lacunae) rather than inventing readings Honest by design: a variant-analysis tool, not a text-retrieval tool — it describes attested readings and names the lexemes at issue; it never presents generated Greek as a codex's diplomatic text Training-knowledge disclaimer banner + "verify against the Göttingen Septuaginta / Rahlfs-Hanhart" guidance on every result SBL footnote + BibTeX export, bilingual (EN/ES) Notes Built on the existing streaming/cache pipeline; pre-seeded featured passages (Judges 5:1, Daniel 4:1, Tobit 1:1, Isaiah 7:14, Genesis 1:1) Code-reviewed clean; honesty behaviors verified against live output BibCrit: An AI-Assisted Web Application for Biblical Textual Criticism BibCrit is an open-source, browser-based toolkit that applies large language model analysis to the methods and problems of biblical textual criticism. It requires no software installation and streams structured scholarly analysis in real time via Server-Sent Events. The application operates fully in English and Spanish, with AI-generated analysis cached and translated server-side. The application integrates fourteen tools in a single interface: MT/LXX Divergence Analyzer — word-level comparison of the Masoretic Text and the Septuagint, classifying each divergence as a translation technique, theological tendency, scribal error, or evidence for a divergent Hebrew Vorlage, with confidence scores and competing scholarly hypotheses. Exports citations in SBL footnote, BibTeX, RIS (Zotero), and TEI XML formats. Back-Translation Workbench — word-by-word retroversion of a LXX passage to its probable Hebrew Vorlage, following Tov's retroversion methodology, with per-word confidence levels and translation strategy annotations. Scribal Tendency Profiler — five-axis radar chart scoring an LXX book's translator on literalness, anthropomorphism reduction, messianic heightening, harmonization, and paraphrase rate, with supporting textual evidence. Numerical Discrepancy Modeler — analysis of numerical divergences between MT, LXX, and the Samaritan Pentateuch (patriarchal ages, census figures, temple dimensions, royal chronology), ranking competing explanations by confidence. Ancient Witness Bridge (DSS) — comparison of a biblical passage across five ancient witnesses: Dead Sea Scrolls manuscripts (1QIsaᵃ, 4QSamᵃ, 11QPaleoLev, 4QDeutᵏ), the Samaritan Pentateuch, the Peshitta (Syriac OT), MT, and LXX. Identifies which witnesses attest the passage, how each aligns with the traditions, and the specific divergences between them. DSS scroll text and Peshitta Syriac are drawn directly from verified corpus data. Theological Revision Detector — identification of potentially theologically motivated alterations including anthropomorphism avoidance, messianic heightening, polemical changes, and harmonization, applying the methodology of Fishbane and Tov. Patristic Citation Tracker — analysis of how Church Fathers through the 5th century cited a passage, what text form their quotations reflect, and what this reveals about early textual transmission. Includes BiblIndex deep-links to primary patristic sources. Manuscript Genealogy — stemmatic visualization of a biblical book's transmission from proto-text through manuscript families (MT, LXX, DSS, SP, Peshitta, Targum, Vulgate) to modern critical editions. NT Use of OT Analyzer — identifies Old Testament allusions and citations in New Testament passages, determines whether the NT author followed the MT or LXX text form, and assesses the degree of verbal correspondence. Methodology: Beale, Hays, Stanley. Chiasm & Literary Structure Detector — detects chiastic and concentric literary structures (A-B-C-B′-A′ patterns), parallel panels, inclusios, and refrains in any biblical passage. Methodology: Lund, Welch, Dorsey, Walsh. Source Criticism Tool — attributes pericopes to documentary sources (J, E, D, P, Redactor) via divine name usage, vocabulary patterns, doublets, and narrative tensions. Methodology: Wellhausen, Friedman, Baden. Targum Comparator — analyses Targum Onkelos (Torah) and Targum Jonathan (Prophets) renderings against the MT and LXX, identifying Memra substitutions, anthropomorphism avoidance, messianic reinterpretation, and targumic expansions. Corpus: Sefaria API, 26 books, 240,297 Aramaic word tokens. Methodology: Samely, Smelik. NT Textual Tradition Analyzer — manuscript family attestation for any New Testament passage (Alexandrian, Western, Byzantine, Caesarean), Metzger A/B/C/D confidence ratings, and a full variant register with patristic support and versional evidence. Methodology: Metzger (1994). Targum Comparator (/targum) — Compare any Torah or Prophets passage against Targum Onkelos (Torah) or Targum Jonathan (Prophets) alongside MT and LXX. Identifies Memra substitutions, anthropomorphism softening, targumic expansions with Midrashic parallels, and messianic reinterpretations. Backed by Sefaria's openly licensed Aramaic corpus (240,297 word tokens, 26 books). NT Textual Tradition Analyzer (/nt-text) — Manuscript family support (Alexandrian, Western, Byzantine, Caesarean), Metzger A/B/C/D confidence ratings, and full variant register for any NT passage. Extended analysis for all major disputed passages (Mark 16:9–20, John 7:53–8:11, 1 John 5:7). Backed by the SBLGNT corpus (137,554 word tokens). Second Temple Literature Bridge (/stl)Maps allusions and parallels between canonical Scripture and five Second Temple works: 1 Enoch, Jubilees, Sirach, 4 Ezra, and Tobit. Classifies each connection by type (direct allusion, thematic parallel, shared tradition), directionality, and confidence, using Nickelsburg / VanderKam / Collins / Knibb / Hays methodology. Corpus and data sources. The application loads eight corpus traditions at startup. The MT corpus uses the ETCBC morphological database (BHSA, Eep Talstra Centre for Bible and Computer, Vrije Universiteit Amsterdam) — all 39 books. The LXX corpus uses Rahlfs-edition data ingested via STEPBible — 38 books. The DSS corpus uses the ETCBC DSS Text-Fabric module — four scrolls: 1QIsaᵃ (Isaiah 1–66), 4QSamᵃ (Samuel fragments), 11QPaleoLev (Leviticus in paleo-Hebrew script), and 4QDeutᵏ (Deuteronomy fragments). The Samaritan Pentateuch uses the dt-ucph SP Text-Fabric module — 5 Torah books. The Peshitta (Syriac OT) uses the ETCBC/peshitta Text-Fabric module (SEDRA, Beth Mardutho) — all 39 OT books, 308,863 word tokens. The Greek New Testament uses MorphGNT / SBLGNT — 27 books. The Targum corpus (Onkelos and Jonathan) uses the Sefaria API — 26 books, 240,297 Aramaic word tokens. The Clementine Vulgate uses the scrollmapper/bible_databases corpus — 66 books, 569,588 Latin word tokens. All corpus data is loaded deterministically at startup with no API calls. Technical stack: Python/Flask backend, Anthropic Claude API (claude-sonnet-4-6) for analysis, Supabase for analysis caching with local JSON fallback, Server-Sent Events for real-time streaming, D3.js for radar chart visualization. All analysis results are cached by SHA-256 key and available via an open data REST API. Bilingual operation (English and Spanish) with server-side translation caching. Released under Apache 2.0.