bible-cpp

Pure C++20 parallel-text scripture engine with plugin architecture.

A zero-dependency, cross-platform library for navigating, searching, and cross-referencing parallel scripture texts. Each translation (KJV, CUV, Synodal, Tao Te Ching, etc.) is a database plugin — the core engine has no hardcoded knowledge of any specific text.

Features

• Plugin architecture — register any parallel text as a ScripturePlugin. Ship KJV, CUV, Russian Synodal, or Tao Te Ching. Add your own.
• Reference parsing — "John 3:16" → (book=43, chapter=3, verse=16). Supports English, Chinese, and extensible locale abbreviations.
• Strong's concordance — Hebrew (H1-H8674) and Greek (G1-G5624) lexicon with lemma, transliteration, morphology, gloss, and full definition. Data from STEPBible (CC BY 4.0).
• Full-text search — FTS5 across all registered parallel texts simultaneously.
• Cross-referencing — tap a Strong's number to find every verse using that word across all translations.
• Statistics — word frequency, Strong's usage counts, per-book/chapter verse counts. AI/MCP-ready metadata on every query result.
• Zero dependencies — pure C++20, no Boost, no nlohmann-json, no std::filesystem. Compiles on 14 platforms via Guix cross-compilation.

Architecture

┌─────────────────────────────────────────┐
│           ScriptureEngine               │
│  navigate, search, reference parsing    │
├─────────────────────────────────────────┤
│         IScripturePlugin                │
│  register text databases as plugins     │
├──────────┬──────────┬───────────────────┤
│   KJV    │  CUV-S   │  CUV-T  │  ...   │
│ (plugin) │ (plugin) │ (plugin)│(plugin) │
└──────────┴──────────┴─────────┴─────────┘
         ↕               ↕
┌─────────────────────────────────────────┐
│         Strong's Lexicon                │
│  Hebrew + Greek concordance data        │
│  (STEPBible TBESG/TBESH, CC BY 4.0)    │
└─────────────────────────────────────────┘

Each plugin is a SQLite database conforming to a standard schema. The engine queries across all registered plugins transparently.

Three-Layer Integration (nomad-builder)

When consumed by nomad-builder, bible-cpp follows the standard three-layer pattern:

Layer	Location	What
1	third_party/bible-cpp/	This repo — standalone MIT library
2	src/libbible/	Nomad wrapper in loqu8::bible namespace
3	src/libbible_capi/	C API for FFI (Flutter, .NET, Electron, Python, WASM)

Data Sources

Dataset	Source	License	Notes
KJV text	Public domain (1611)	Public Domain	King James Version
CUV text	Public domain (1919)	Public Domain	Chinese Union Version (simplified + traditional)
Strong's Concordance	James Strong (1890)	Public Domain	Original concordance
Strong's Hebrew Lexicon (TBESH)	STEPBible	CC BY 4.0	Tyndale House, Cambridge
Strong's Greek Lexicon (TBESG)	STEPBible	CC BY 4.0	Tyndale House, Cambridge
CUV Strong's alignment	Bible SuperSearch	Public Domain (text)	Alignment-based tagging

Attribution: Strong's lexicon data from "Tyndale House, Cambridge" (tyndaleHouse.com) and "STEP Bible" (STEPBible.org). Source: STEPBible-Data.

Database Schema

Plugin Schema (per-translation database)

CREATE TABLE books (
    id INTEGER PRIMARY KEY,           -- 1-66 for Protestant Bible
    name TEXT NOT NULL,                -- primary name (language of this plugin)
    name_alt TEXT,                     -- alternate name
    abbr TEXT NOT NULL,                -- abbreviation for reference parsing
    testament TEXT CHECK(testament IN ('OT', 'NT', NULL)),
    chapter_count INTEGER NOT NULL
);

CREATE TABLE verses (
    id INTEGER PRIMARY KEY AUTOINCREMENT,
    book INTEGER NOT NULL REFERENCES books(id),
    chapter INTEGER NOT NULL,
    verse INTEGER NOT NULL,
    text TEXT,                         -- verse text (may contain Strong's markers)
    text_plain TEXT,                   -- stripped text (no markers)
    UNIQUE(book, chapter, verse)
);

CREATE INDEX idx_verses_book_ch ON verses(book, chapter);

Strong's Lexicon Schema

CREATE TABLE strongs (
    number INTEGER NOT NULL,
    type TEXT NOT NULL CHECK(type IN ('H', 'G')),
    language TEXT NOT NULL DEFAULT 'Hebrew'
        CHECK(language IN ('Hebrew', 'Aramaic', 'Greek')),
    lemma TEXT,                        -- original word (אֱלֹהִים / λόγος)
    translit TEXT,                     -- transliteration
    pronunciation TEXT,
    morph TEXT,                        -- morphological classification
    gloss TEXT,                        -- one-word gloss
    definition TEXT,                   -- full lexical definition
    PRIMARY KEY(number, type)
);

CREATE INDEX idx_strongs_type ON strongs(type);

Adding a New Translation

To add a Russian Synodal Bible (or any other text):

1. Create a SQLite database conforming to the plugin schema above
2. Populate books with Russian book names and abbreviations
3. Populate verses with the text (optionally with {H1234} Strong's markers)
4. Register the plugin:

#include <bible/bible.hpp>

bible::ScriptureEngine engine;
engine.register_plugin("synodal", "/path/to/synodal.sqlite", {
    .language = "ru",
    .name = "Synodal Translation",
    .has_strongs = false
});

The engine handles navigation, search, and cross-referencing automatically.

Building

mkdir build && cd build
cmake .. -DCMAKE_BUILD_TYPE=Release
cmake --build .
ctest

Requirements

• C++20 compiler (GCC 12+, Clang 15+, MSVC 2022+)
• CMake 3.20+
• No external dependencies (SQLite is vendored or expected via find_package)

MCP / AI Agent Integration

bible-cpp is designed for AI agent consumption. Every query returns structured metadata:

auto result = engine.search("love", {.limit = 10});
// result.stats.query_time_ms, result.stats.total_matches
// result.stats.books_matched, result.stats.translations_searched

auto stats = engine.compute_stats();
// stats.total_verses, stats.total_books, stats.plugins_registered
// stats.strongs_entries, stats.search_index_size

When wrapped in an MCP server (Layer 3 CAPI → MCP tools), this enables:

MCP Tool	What It Does
bible_get_books	List all books with chapter counts
bible_get_chapter	Get full chapter across all translations
bible_get_verse	Get single verse with all translations + Strong's
bible_search	Full-text search across all registered texts
bible_get_strongs	Look up Strong's concordance entry
bible_get_stats	Corpus statistics, word frequency, coverage
bible_navigate	Navigate UI to specific passage
bible_parse_ref	Parse "John 3:16" → structured coordinates

Project Status

Phase: Engine complete, data pipeline operational. Next: nomad-builder integration.

• Architecture design (plugin model, schema, CAPI pattern)
• Core types and interfaces (types.hpp, scripture_plugin.hpp)
• Reference parser (reference.hpp/cpp) — English, Chinese, abbreviations
• SQLite query layer (navigation, search, parallel text)
• Strong's lexicon integration (STEPBible TBESG/TBESH)
• Build tools (tools/build_bible_db.py — auto-downloads + builds 4 databases)
• Statistics / analytics (AI/MCP-ready corpus metadata)
• Tests (37 passing — strongs parser + full engine integration)
• nomad-builder integration (Layer 2 + Layer 3 CAPI)

License

Code: MIT License — see LICENSE.

Data: Strong's lexicon data is CC BY 4.0 from STEPBible-Data by Tyndale House, Cambridge. Bible texts (KJV, CUV) are public domain.