call.call()

Run one CLI-backed LLM call and parse its response.

Usage

Source

call.call(
    prompt,
    *,
    backend,
    model="small",
    agent=False,
    response_model=None,
)

Parameters

prompt: str

The user prompt, delivered to the backend over stdin.

backend: LlmBackend

The LlmBackend to invoke.

model: TModel = "small"

Abstract model tier (small/medium/large).

agent: bool = False

Whether the call may use tools / agent capabilities.

response_model: type[BaseModel] | None = None

Pydantic model for structured output, or None for text.

Returns

str | BaseModel

The raw text response, or a validated response_model instance.