Usage - CosVec

Calling the System One API

Async
Sync

import asyncio

from typesafe_sdk import AsyncTypeSafeClient, Choice, Noul, Score


async def main() -> None:
    async with AsyncTypeSafeClient() as client:
        result = await client.system_one(
            "I was charged twice. Please help ASAP.",
            {
                "billing": Noul(instructions="Is this about billing?"),
                "tone": Choice(
                    instructions="What is the tone?",
                    criteria={"calm": None, "angry": None},
                ),
                "urgency": Score(
                    instructions="How urgent is this?",
                    criteria=["low", "medium", "high"],
                ),
            },
        )
        print(
            result.nouls["billing"].noul,
            result.choices["tone"].choice,
            result.scores["urgency"].score,
        )


asyncio.run(main())

from typesafe_sdk import Choice, Noul, Score, TypeSafeClient

client = TypeSafeClient()
state = "I was charged twice. Please help ASAP."
questions = {
    "billing": Noul(instructions="Is this about billing?"),
    "tone": Choice(
        instructions="What is the tone?", criteria={"calm": None, "angry": None}
    ),
    "urgency": Score(
        instructions="How urgent is this?", criteria=["low", "medium", "high"]
    ),
}
result = client.system_one(state, questions)
print(
    result.nouls["billing"].noul,
    result.choices["tone"].choice,
    result.scores["urgency"].score,
)

Typed `system_one` responses

It is possible to provide a response model to system_one to make using the response more type-safe:

from typesafe_sdk import Noul, NoulAnswer, SystemOneResponse, TypeSafeClient


class BillingResponse(SystemOneResponse):
    billing: NoulAnswer


with TypeSafeClient() as client:
    result = client.system_one(
        "I was charged twice.",
        {"billing": Noul(instructions="Is this about billing?")},
        response_model=BillingResponse,
    )
    assert 0 <= result.billing.noul <= 1
    assert result.billing == result.nouls["billing"]
    print(result.request_id)

Custom response types

It is also possible to define a completely new response model without inheriting from SystemOneResponse:

from pydantic import BaseModel

from typesafe_sdk import Noul, NoulAnswer, TypeSafeClient


class BillingAnswers(BaseModel):
    billing: NoulAnswer


class BillingResponse(BaseModel):
    answers: BillingAnswers


result = TypeSafeClient().system_one(
    "I was charged twice.",
    {"billing": Noul(instructions="Is this about billing?")},
    response_model=BillingResponse,
)
assert 0 <= result.answers.billing.noul <= 1

Choosing a model

Inspect the available models:

from typesafe_sdk import TypeSafeClient

print(TypeSafeClient().models.list())

Select the model when constructing a client:

client = TypeSafeClient(model="jev")

See the Models resource reference for details.

Retries

Pass a custom RetryPolicy as retry on the client or per call.

Client
Per-call

from typesafe_sdk import RetryPolicy, TypeSafeClient

client = TypeSafeClient(retry=RetryPolicy(max_retries=3, backoff_max=0.2, timeout=1.0))

from typesafe_sdk import RetryPolicy

client.system_one(
    state, questions, retry=RetryPolicy(max_retries=3, backoff_max=0.2, timeout=1.0)
)

Error handling

Handle exceptions raised by the SDK:

from typesafe_sdk import TypeSafeAPIError

try:
    client.system_one(state, questions)
except TypeSafeAPIError as error:
    print(error.status, error.request_id)

Logging

The SDK logs to the typesafe_sdk logger. Configure it according to standard logging guide:

import logging

logging.getLogger("typesafe_sdk").setLevel(logging.DEBUG)

Or set TYPESAFE_LOG_LEVEL to one of debug, info, warning, error, or off before importing the SDK. info logs one summary line per request; debug also logs request and response headers and bodies. Secret headers — authorization, API keys, cookies, and any header whose name contains token or secret — are redacted from log output. Request and response bodies are not redacted.

Environment variables

The SDK reads and uses the following environment variables:

Variable	Configures	Default
`TYPESAFE_API_KEY`	API key (required)	—
`TYPESAFE_BASE_URL`	API root URL	`https://api.typesafe.ai`
`TYPESAFE_DEFAULT_MODEL`	Default model	`jev-latest`
`TYPESAFE_LOG_LEVEL`	`typesafe_sdk` logger level, applied once at import	unset

See the constants reference for SDK defaults.

Forward compatibility

The SDK keeps working as the CosVec API evolves, so you can adopt new API features before an SDK release adds first-class support for them.

Extra request fields

Send request fields this SDK version predates with extra_body:

from typesafe_sdk import Noul, TypeSafeClient

with TypeSafeClient() as client:
    client.system_one(
        "I was charged twice.",
        {"billing": Noul(instructions="About billing?")},
        extra_body={"beam_width": 4},
    )

Raw question dictionaries

from typesafe_sdk import TypeSafeClient

with TypeSafeClient() as client:
    client.system_one(
        "I was charged twice.",
        {"billing": {"type": "noul", "instructions": "About billing?", "weight": 2}},
    )

TipUnknown fields are a forward-compatibility escape hatch. Ignore their type-checking errors and prefer upgrading the SDK instead.

Unknown answer kinds

The SDK logs a warning and skips unrecognized answer kinds. Use raw_http_response to inspect the complete API response, including those answers:

from typesafe_sdk import Noul, TypeSafeClient

result = TypeSafeClient().system_one(
    "I was charged twice.",
    {"billing": Noul(instructions="Is this about billing?")},
)
raw_answers = result.raw_http_response.json()["answers"]

Unknown response fields

Unknown extra fields on recognized responses are ignored.

​Calling the System One API

​Typed system_one responses

​Custom response types

​Choosing a model

​Retries

​Error handling

​Logging

​Environment variables

​Forward compatibility

​Extra request fields

​Raw question dictionaries

​Unknown answer kinds

​Unknown response fields

Calling the System One API

Typed `system_one` responses

Custom response types

Choosing a model

Retries

Error handling

Logging

Environment variables

Forward compatibility

Extra request fields

Raw question dictionaries

Unknown answer kinds

Unknown response fields