Extract

See how to use extract() to extract structured data from web pages

​
Method Signatures

// With schema and options
await page.extract<T extends z.AnyZodObject>(options: ExtractOptions<T>): Promise<ExtractResult<T>>

// String instruction only
await page.extract(instruction: string): Promise<{ extraction: string }>

// No parameters (raw page content)
await page.extract(): Promise<{ page_text: string }>
ExtractOptions Interface:
interface ExtractOptions<T extends z.AnyZodObject> {
  instruction?: string;
  schema?: T;
  modelName?: AvailableModel;
  modelClientOptions?: ClientOptions;
  domSettleTimeoutMs?: number;
  selector?: string;
  iframes?: boolean;
}

type ExtractResult<T> = z.infer<T>;

​
Parameters

​
instruction
string
Natural language description of what data to extract.
​
schema
z.ZodSchema | BaseModel
Type schema defining the structure of data to extract. Ensures type safety and validation.
​
selector
string
XPath selector to limit extraction scope. Reduces token usage and improves accuracy.
​
iframes
boolean
Set to true if content exists within iframes.Default: false
​
modelName
AvailableModel
Override the default LLM model for this extraction.
​
modelClientOptions
ClientOptions
Model-specific configuration options.
​
domSettleTimeoutMs
number
Maximum time to wait for DOM to stabilize.Default: 30000

​
Response Types

Returns: Promise<ExtractResult<T>> where T matches your schemaThe returned object will be strictly typed according to your schema definition.

​
Code Examples

import { z } from 'zod';

// Schema definition
const ProductSchema = z.object({
  name: z.string(),
  price: z.number(),
  inStock: z.boolean()
});

// Extraction
const product = await page.extract({
  instruction: "extract product details",
  schema: ProductSchema
});

​
Example Response

{
  "name": "Product Name",
  "price": 100,
  "inStock": true
}