BibleQL Python

A typed Python client for the BibleQL GraphQL API. Query Bible verses, passages, translations, search and concordance across 47 translations in 31 languages, without writing GraphQL.

Installation

pip install bibleql

Requires Python 3.10+.

Quick start

import bibleql

with bibleql.Client(api_key="bql_live_...") as client:
    passage = client.passage("John 3:16")
    print(passage.reference)  # John 3:16
    print(passage.text)  # For God so loved the world...

Get an API key at bibleql.org, and try queries in the playground.

Configuration

client = bibleql.Client(
    api_key="bql_live_...",  # or set the BIBLEQL_API_KEY environment variable
    default_translation="spa-bes",  # optional, defaults to "eng-web"
    api_url="https://custom-host.com/graphql",  # optional, defaults to https://bibleql.org/graphql
    timeout=60,  # optional, seconds, defaults to 30
)

If api_key is omitted, the client reads BIBLEQL_API_KEY:

export BIBLEQL_API_KEY=bql_live_...

client = bibleql.Client()

Use the client as a context manager (with ... as client:) or call client.close() when done, so its connection pool is released.

Every method that takes translation= falls back to default_translation when it's omitted. The exception is semantic_search, which lets the API pick a translation that has embeddings.

Usage

Passages

passage = client.passage("John 3:16")
passage.reference  # "John 3:16"
passage.text  # "For God so loved the world..."
passage.translation_id  # "eng-web"
passage.verses  # [Verse(book_id='JHN', book_name='John', chapter=3, verse=16, text='...')]

client.passage("John 3:16-18")
client.passage("Matthew 25:31-33,46")
client.passage("Genesis 1")
client.passage("Mateo 28:18-20", translation="spa-bes")

Verses and chapters

verse = client.verse("MAT", 5, 3)
verse.book_name  # "Matthew"
verse.text  # "Blessed are the poor in spirit..."

verses = client.chapter("GEN", 1)  # list[Verse]

Random verse

client.random_verse()
client.random_verse(testament="NT")  # "OT" or "NT"
client.random_verse(books=["PSA", "PRO"])  # or books="PSA,PRO"

Verse of the day

import datetime

client.verse_of_the_day()
client.verse_of_the_day(date=datetime.date(2026, 1, 1))  # or date="2026-01-01"

Search

Case-insensitive substring search, with no stemming ("love" won't match "loved"):

for verse in client.search("love", limit=5):
    print(verse.book_name, verse.chapter, verse.verse, verse.text)

Semantic search finds verses by meaning, using AI embeddings:

for result in client.semantic_search("fe y esperanza", limit=5):
    print(f"{result.similarity:.2f}", result.verse.text)

Concordance

Exhaustive, canonically ordered occurrences of a word, with per-book and per-testament totals:

page = client.concordance("misericordia", translation="spa-rv1909", first=10)
page.total_count  # total occurrences
page.entry.surface_forms  # ["misericordia", "misericordias", ...]
page.entry.occurrences_by_testament.old  # verses in the Old Testament
for occurrence in page.occurrences:
    print(occurrence.verse.book_name, occurrence.context)  # context has <mark> tags

# Next page
if page.has_next_page:
    page = client.concordance("misericordia", translation="spa-rv1909", after=page.end_cursor)

# Or let the client paginate for you (pages are fetched lazily)
for occurrence in client.iter_concordance("misericordia", translation="spa-rv1909"):
    ...

# Filters
client.concordance("misericordia", translation="spa-rv1909", book="Salmos")
client.concordance("misericordia", translation="spa-rv1909", testament=bibleql.Testament.NEW)

# Alphabetical word-frequency index
client.concordance_index(translation="spa-rv1909", prefix="mis", min_occurrences=5)

The context field contains <mark> HTML tags around the match, so sanitize it before rendering it as HTML.

Translations, languages, books

client.translations()  # list[Translation]
t = client.translation("eng-web")
t.name, t.abbrev  # "World English Bible", "WEB"
t.books  # list[LocalizedBook]

client.languages()  # list[Language], each with .translations
client.books()  # list[Book], the 66 canonical books

Bible index

index = client.bible_index()
index[0].name  # "Genesis"
index[0].chapters[0].verse_count  # 31

Models

All results are immutable dataclasses with snake_case attributes. To convert one to a dict:

import dataclasses

dataclasses.asdict(client.verse("JHN", 3, 16))

Error handling

Every error inherits from bibleql.BibleQLError:

BibleQLError
├── ConfigurationError        no API key
├── NetworkError              no response (DNS, connection refused, ...)
│   └── RequestTimeoutError
├── APIError (.status, .body) non-2xx HTTP response
│   ├── AuthenticationError     401
│   ├── RateLimitError          429
│   └── ServerError             5xx
└── QueryError (.errors)      GraphQL errors in the response
    └── NotFoundError

try:
    client.passage("Hezekiah 1:1")
except bibleql.NotFoundError as e:
    print(f"Not found: {e}")
except bibleql.AuthenticationError:
    print("Invalid API key")
except bibleql.RateLimitError:
    print("Too many requests")
except bibleql.RequestTimeoutError:
    print("Request timed out")
except bibleql.NetworkError:
    print("Connection failed")
except bibleql.QueryError as e:
    print(f"GraphQL error: {e}")
    print(e.errors)  # raw error list from the API

Development

asdf install                      # Python version from .tool-versions
python -m venv .venv && source .venv/bin/activate
python -m pip install -e ".[dev]"

pytest                            # run tests (HTTP is mocked with respx, no network needed)
ruff check . && ruff format .     # lint and format

tests/fixtures/schema.graphql is a snapshot of the API schema, and every query is validated against it. Refresh it from the API repo (docs/generated/schema.graphql) when the schema changes.

Testing against the live API

The test suite never touches the network. To check the client against the real API, put your key in a .env file (gitignored) and run the smoke test, which loads it with python-dotenv:

cp .env.example .env              # then set BIBLEQL_API_KEY in .env
python scripts/smoke_test.py      # calls every method and prints the results

The library itself never reads .env files. In your own app, load them yourself (for example with dotenv.load_dotenv()) before creating a Client().

License

Released under the MIT License.