Tomopas/bibleai-audio

Bible AI — King James Version Audio

Chapter-by-chapter narration of the complete King James Version (KJV) Bible, generated locally with Kokoro-82M.

Files	1,189 MP3s (66 books, one file per chapter)
Size	About 2.14 GB
Format	MP3, 24 kHz mono, about 64 kbps
Voice	af_heart (American English)
TTS model	hexgrad/Kokoro-82M (Apache-2.0)
Source text	King James Version (public domain in the United States)
Generated	2026-10-26 (UTC)
Companion code	tomopas/BibleAIApp

Layout

audio/en-kjv/{book}/{chapter}.mp3

book is a lowercase three-letter book code, and chapter is the chapter number without leading zeros. For example:

audio/en-kjv/gen/1.mp3
audio/en-kjv/psa/119.mp3
audio/en-kjv/jhn/3.mp3

audio/en-kjv/kokoro-progress.json records generation metadata, including the model, voice, and per-chapter completion times.

Book codes

Code	Book
gen	Genesis
exo	Exodus
lev	Leviticus
num	Numbers
deu	Deuteronomy
jos	Joshua
jdg	Judges
rut	Ruth
1sa	1 Samuel
2sa	2 Samuel
1ki	1 Kings
2ki	2 Kings
1ch	1 Chronicles
2ch	2 Chronicles
ezr	Ezra
neh	Nehemiah
est	Esther
job	Job
psa	Psalms
pro	Proverbs
ecc	Ecclesiastes
sng	Song of Songs
isa	Isaiah
jer	Jeremiah
lam	Lamentations
eze	Ezekiel
dan	Daniel
hos	Hosea
joe	Joel
amo	Amos
oba	Obadiah
jon	Jonah
mic	Micah
nah	Nahum
hab	Habakkuk
zep	Zephaniah
hag	Haggai
zec	Zechariah
mal	Malachi
mat	Matthew
mrk	Mark
luk	Luke
jhn	John
act	Acts
rom	Romans
1co	1 Corinthians
2co	2 Corinthians
gal	Galatians
eph	Ephesians
php	Philippians
col	Colossians
1th	1 Thessalonians
2th	2 Thessalonians
1ti	1 Timothy
2ti	2 Timothy
tit	Titus
phm	Philemon
heb	Hebrews
jas	James
1pe	1 Peter
2pe	2 Peter
1jn	1 John
2jn	2 John
3jn	3 John
jud	Jude
rev	Revelation

Usage

Download one chapter:

from huggingface_hub import hf_hub_download

path = hf_hub_download(
    repo_id="Tomopas/bibleai-audio",
    filename="audio/en-kjv/gen/1.mp3",
    repo_type="dataset",
)

Download the whole dataset:

hf download Tomopas/bibleai-audio --repo-type dataset --local-dir bibleai-audio

Checksums

audio-manifest.sha256 lists the SHA-256 of every MP3 using the same paths as the dataset. From the dataset root, verify a download with:

sha256sum -c audio-manifest.sha256

Generation

Audio was generated from the KJV text with Kokoro's local TTS pipeline, one MP3 per chapter, and encoded with ffmpeg. No hosted TTS provider or web research was used. The reproducible command in the companion repository is:

python worker/generate_kokoro_audio.py --run --translation en-kjv --yes

License and attribution

• The generated narration is released under CC BY 4.0. See DATA_LICENSE.md.
• The KJV source text is public domain in the United States. It is subject to perpetual Crown copyright in the United Kingdom; check the law in your jurisdiction.
• Kokoro-82M is licensed under Apache-2.0. Model licenses do not automatically determine ownership of generated output.

Suggested attribution:

Bible AI generated KJV narration, generated with Kokoro-82M voice af_heart. CC BY 4.0.

Limitations

• Single voice, read in chapter order. There is no verse-level alignment or timestamp metadata.
• TTS may mispronounce names, numbers, and archaic terms.
• This is a generated reading aid, not a critical edition or a substitute for a verified Bible text. It is not affiliated with or endorsed by any church or Bible publisher.