auray-client

The Auray API from Python. One dependency, httpx; Python 3.10+.

from auray_client import Client

auray = Client()  # reads AURAY_API_KEY

result = auray.subscribe(
    "auray-ai/krea-2-turbo/text-to-image",
    input={"prompt": "a brass orrery on a walnut desk, morning light", "aspect_ratio": "16:9", "resolution": 1024},
    idempotency_key="my-first-render",
    on_queue_update=lambda s: print(s["status"]),
)
assets = auray.platform.jobs.assets(result["request_id"])
print(assets["assets"][0]["url"])

• run(address, input=..., wait_seconds=..., timeout=...) — the synchronous host; a 504 run_timed_out is followed by the job it names (queue_request_id, never the trace in request_id), never resubmitted. wait_seconds is how long the host holds the connection, timeout the whole call's deadline; running out raises subscribe_timed_out carrying the job.
• subscribe(address, input=..., idempotency_key=..., on_queue_update=...) — submit, follow, fetch the result.
• queue.submit / status / result / cancel / stream_status — the request lifecycle, fal's shape.
• storage.upload(product=..., content_type=..., data=...) and upload_file(path, product=...) — three calls, and a signed URL for a field.
• platform.me / usage / jobs / models / webhooks / chat — the account's side.
• AsyncClient — the same surface, awaited: async with AsyncClient() as auray: ...

Every refusal raises AurayError with status, code, the API's message, docs and request_id. 429 and 503 are retried with Retry-After; nothing else is.

Documentation: https://auray.ai/docs/clients/python

Publishing, and why there is no build script

Publishing to PyPI is the owner's decision and has not been taken, so nothing here builds a distribution. The package carried "build": "python -m build" from the day it landed and it never ran: build is not in requirements-dev.txt, is not installed on the machine this repository is developed on, and is not installed in CI — where pnpm turbo run typecheck lint test build runs on every push, so the script turned the whole pipeline red from that day to 2026-09-06, next to a generated-sdks job that was red for its own reason. A red pipeline that everybody has learned to ignore is worse than no pipeline, so the script is gone until the decision is made.

When it is: pip install build, add it to requirements-dev.txt, python -m build (the backend is already declared — hatchling, in pyproject.toml), and give CI the install step before the build task exists again.