Skip to main content
agent.extract(query: str, schema: dict, model: str | None = None) -> dict | list
Returns data matching a JSON Schema.

​
Parameters

ParameterTypeDefaultDescription
querystr—What data to extract from the screen
schemadict—JSON Schema defining the output format
modelstr | NoneNoneOverride the default model for this call

​
Returns

dict | list — Structured data matching the schema.

​
Raises

  • WorkflowError — on extraction failure
  • ValueError — if schema is empty

​
Examples

Inline schema: 
def run(params: Params) -> Result:
    agent = Agent()
    data = agent.extract(
        "What is the page title?",
        {"type": "object", "properties": {"title": {"type": "string"}}, "required": ["title"]}
    )
Using Pydantic models (recommended): 
from pydantic import BaseModel

class ProductInfo(BaseModel):
    name: str
    price: float

def run(params: Params) -> Result:
    agent = Agent()
    data = agent.extract("Extract the product details", ProductInfo.model_json_schema())
    product = ProductInfo.model_validate(data)
With model override for complex extraction: 
def run(params: Params) -> Result:
    agent = Agent()
    data = agent.extract(
        "Extract all visible invoice line items",
        InvoiceSchema.model_json_schema(),
        model="claude-sonnet-4-5-20250929"
    )
Use YourModel.model_json_schema() to generate the schema from your Pydantic models — this keeps your schema in sync with your types.