The API

Why this exists

If you have tried to put a modern Bible translation in an app, you already know the problem. Permission requests that take eight months, or a year and a half, or that are never answered. Forms asking for your distribution and marketing strategy. Licences that can be withdrawn after you ship, leaving you to strip the text out and absorb the reviews. Terms that forbid you from publicly criticising a rendering.

The free alternatives solve the legal problem and create a different one: they are all written in church English. If you are building for people with no church background, for teenagers, for prisons, for recovery groups, for anyone learning the language, your free options sound like a hymnal and your plain options need a licence.

This is a plain-English translation under an open licence. Both problems, one answer.

The licence, in full

Creative Commons Attribution 4.0. Credit the translation and you are finished. In particular, and because these are the questions people actually have to ask a rights desk about other translations:

• ✓Commercial use, including paid apps and subscriptions.
• ✓Storing the text as embeddings in a vector database.
• ✓Including it in a model training or fine-tuning corpus.
• ✓A model reproducing it verbatim in its output.
• ✓Printing, including print-on-demand and printing at home.
• ✓Reading it aloud on a podcast, in a broadcast, or as an audio Bible.
• ✓Publicly criticising any rendering in it.

There is no verse ceiling, no percentage cap, no whole-book restriction, and no renewal. The licence is irrevocable. Machine-readable terms live at GET /api/v1/license.

Get a key

No approval step. The key works on the response to this request.

curl -X POST https://ngetbible.com/api/v1/keys \
  -H 'Content-Type: application/json' \
  -d '{"email":"you@example.com","name":"what you are building"}'

We ask for an email for one reason: so we can warn you before anything changes. That is the promise the rest of this page is built on, and it is worth nothing if we cannot reach you.

Limits

Every response carries X-RateLimit-Remaining and X-RateLimit-Reset. Partner access is free; it exists so that someone embedding the whole Bible does not have to pretend to be several people. Reply to the email your key came from and say what you are doing.

Pace your requests

One practical warning, because it will otherwise cost you an afternoon. This API runs behind a platform with its own automatic denial-of-service protection, and that protection sits in front of our code. If you fire requests as fast as your machine can open sockets, you can trip it well before you reach the limits above, and it answers with an HTML page rather than JSON.

If you get back HTML containing Security Checkpoint, or a 403 with no X-RateLimit headers on it, that is what happened. It is not your key and it is not a ban. Space your requests out and it clears.

A steady two or three requests a second is comfortable. If you are backfilling a database, use /api/v1/export instead of walking chapters in a tight loop; that is what it is for, and one call gets you a whole book.

Endpoints

A first request

curl -G https://ngetbible.com/api/v1/lookup \
  --data-urlencode 'ref=John 3:16; 1 Cor 13:4-7' \
  -H 'Authorization: Bearer nget_your_key_here'

And the part that is worth the trip:

curl https://ngetbible.com/api/v1/concepts/exodus/3?verses=5 \
  -H 'Authorization: Bearer nget_your_key_here'

That returns the Hebrew behind “the ground you’re standing on belongs to God”, and the reasoning for not printing the word every other Bible prints there.

Building a corpus

One book at a time, paginated by the header, licence attached to the file:

curl 'https://ngetbible.com/api/v1/export?book=psalms&format=jsonl' \
  -H 'Authorization: Bearer nget_your_key_here' -o psalms.jsonl

Or take the whole thing with book=all, which walks Genesis 1:1 to Revelation 22:21 in canonical order. Follow X-NGET-Next-Offset until X-NGET-Has-More is false. Four pages covers all 31,103 verses.

OFFSET=0
while [ -n "$OFFSET" ]; do
  curl -sD /tmp/h "https://ngetbible.com/api/v1/export?book=all&offset=$OFFSET" \
    -H 'Authorization: Bearer nget_your_key_here' >> bible.jsonl
  OFFSET=$(grep -i x-nget-next-offset /tmp/h | tr -d '\r' | cut -d' ' -f2)
done

The concept notes are opt-in on a whole-Bible export, with &concepts=true, because they are 30 MB against 3.8 MB for the text alone. Pages get smaller when you ask for them.