tubox-veyra 1.0.1

The official Veyra Python SDK — sync + async, fully typed.

veyra-python

The official Veyra Python SDK: sync + async clients, typed models, streaming, pagination, retries, and rich API exceptions.

Installation

pip install veyra

Quickstart (Sync)

import veyra

client = veyra.Veyra()
completion = client.chat.completions.create(
    model="gpt-5.4-mini",
    messages=[{"role": "user", "content": "Hello!"}],
    reasoning={"effort": "low"},
)
print(completion.choices[0].message.content)

Quickstart (Async)

import asyncio
import veyra

async def main() -> None:
    async with veyra.AsyncVeyra() as client:
        completion = await client.chat.completions.create(
            model="gpt-5.4-mini",
            messages=[{"role": "user", "content": "Hello from async"}],
        )
        print(completion.choices[0].message.content)

asyncio.run(main())

Authentication

Set environment variable:

export VEYRA_API_KEY="veyra_sk_..."

Or pass explicitly:

client = veyra.Veyra(api_key="veyra_sk_...")

Resources at a glance

Namespace	Method
client.chat.completions	create(...)
client.completions	create(...)
client.responses	create(...)
client.embeddings	create(...)
client.images.generations	create(...)
client.audio.transcriptions	create(...)
client.models	list(), retrieve(id)
client.quota	status(), list_plans(), list_public_plans()
client.billing.usage	list(), daily_summary(), monthly_summary(), conversation_summary(id)
client.billing.profile	retrieve(), upsert(...), access()
client.api_keys	create(...), list(), update(...), revoke(...)
client.assistant	chat(...)
client.health	check(), ready()

Streaming

import veyra

client = veyra.Veyra()
with client.chat.completions.create(
    model="gpt-5.4-mini",
    messages=[{"role": "user", "content": "Count to 3"}],
    stream=True,
) as stream:
    for chunk in stream:
        print(chunk.choices[0].delta.content or "", end="")

Responses reasoning

response = client.responses.create(
    model="gpt-5.4-mini",
    input="Summarize the tradeoff in one sentence.",
    reasoning={"effort": "medium", "summary": "auto"},
    max_output_tokens=128,
)
print(response.output[0].content[0].text)

Error handling

import veyra

client = veyra.Veyra()
try:
    client.chat.completions.create(
        model="gpt-5.4-mini",
        messages=[{"role": "user", "content": "Hello"}],
    )
except veyra.RateLimitError as exc:
    print("Retry after:", exc.retry_after)
except veyra.APIError as exc:
    print(exc.status_code, exc.code, exc)

Pagination

for record in client.billing.usage.list(limit=100):
    print(record.model, record.total_tokens)

Retries and timeouts

client = veyra.Veyra(max_retries=4, timeout=30)
client.chat.completions.create(
    model="gpt-5.4-mini",
    messages=[{"role": "user", "content": "Hello"}],
    timeout=10,
)

Raw responses

raw = client.with_raw_response.chat.completions.create(
    model="gpt-5.4-mini",
    messages=[{"role": "user", "content": "Hi"}],
)
print(raw.request_id)
print(raw.http_status)

Async streaming

import asyncio
import veyra

async def main() -> None:
    async with veyra.AsyncVeyra() as client:
        stream = await client.chat.completions.create(
            model="gpt-5.4-mini",
            messages=[{"role": "user", "content": "Count to 3"}],
            stream=True,
        )
        async with stream:
            async for chunk in stream:
                print(chunk.choices[0].delta.content or "", end="")

asyncio.run(main())

Full docs: docs/