Structured output - swarms

Pass a pydantic model as response_format to get JSON that matches it:

from pydantic import BaseModel

import routehub


class RiskReport(BaseModel):
    title: str
    severity: int


response = routehub.completion(
    model="gpt-5.4-mini",
    messages=[{"role": "user", "content": "Assess the main risk of a single-region deployment."}],
    response_format=RiskReport,
)
report = RiskReport.model_validate_json(
    response.choices[0].message.content
)
print(report.title, report.severity)

How it works

A pydantic class becomes a strict json_schema response format.
A dict is sent as given, so you can pass {"type": "json_object"} or your own json_schema block.
The JSON arrives as the message content string. Parse it yourself, as above.
For Claude, RouteHub enforces the schema through a forced tool call, or native structured output on Claude 5.1 and later, and still returns the result as the message content.

The same code works on every provider that supports structured output:

for model in ["gpt-5.4-mini", "claude-sonnet-4-6", "gemini/gemini-3-flash-preview"]:
    response = routehub.completion(
        model=model, messages=messages, response_format=RiskReport
    )
    print(model, RiskReport.model_validate_json(response.choices[0].message.content))

routehub.supports_response_schema(model) tells you whether the catalog lists structured-output support for a model. See Model catalog.