Skip to main content
POST
/
scrape
python
from notte_sdk import NotteClient
from pydantic import BaseModel


class CompanyInfo(BaseModel):
    name: str
    foundation_year: int
    location: str


notte = NotteClient()
company_info = notte.scrape(
    url="https://www.ycombinator.com/companies/nottelabs",
    instructions="Extract the company info from the webpage",
    response_format=CompanyInfo,
    only_main_content=False,
)
{
  "markdown": "<string>",
  "images": [
    {
      "url": "<string>",
      "description": "<string>"
    }
  ],
  "structured": {
    "success": true,
    "error": "<string>",
    "data": {}
  }
}

Documentation Index

Fetch the complete documentation index at: https://docs.notte.cc/llms.txt

Use this file to discover all available pages before exploring further.

Authorizations

​
Authorization
string
header
required

The access token received from the authorization server in the OAuth 2.0 flow.

Headers

​
x-notte-request-origin
string | null
​
x-notte-sdk-version
string | null

Body

application/json
​
url
string
required
​
headless
boolean
default:true

Whether to run the session in headless mode.

​
solve_captchas
boolean
default:false

Whether to try to automatically solve captchas

​
max_duration_minutes
integer
default:15

Maximum session lifetime in minutes (absolute maximum, not affected by activity).

Required range: x <= 1440
​
idle_timeout_minutes
integer
default:3

Idle timeout in minutes. Session closes after this period of inactivity (resets on each operation).

Required range: x <= 15
​
proxies

List of custom proxies to use for the session. If True, the default proxies will be used.

​
browser_type
enum<string>
default:chromium

The browser type to use. Can be chromium, chrome or firefox.

Available options:
chromium,
chrome,
firefox,
chrome-nightly,
chrome-turbo
​
user_agent
string | null

The user agent to use for the session

​
chrome_args
string[] | null

Overwrite the chrome instance arguments

​
viewport_width
integer | null

The width of the viewport

​
viewport_height
integer | null

The height of the viewport

​
aspect_ratio
enum<string> | null

Viewport shape preset. When set, the backend fits the largest rectangle of this aspect ratio inside the sampled available screen area. Cannot be combined with explicit viewport_width/viewport_height.

Available options:
5:4,
16:9
​
cdp_url
string | null

The CDP URL of another remote session provider.

​
use_file_storage
boolean
default:true

Whether FileStorage should be attached to the session.

​
screenshot_type
enum<string>
default:last_action

The type of screenshot to use for the session.

Available options:
raw,
full,
last_action
​
profile
SessionProfile · object

Browser profile configuration for state persistence

​
web_bot_auth
boolean
default:false

Whether to use web bot authentication.

​
extra_http_headers
Extra Http Headers · object

Extra HTTP headers to be sent with every request.

​
vault_id
string | null

The vault to use for the session

​
selector
string | null

Playwright selector to scope the scrape to. Only content inside this selector will be scraped.

​
scrape_links
boolean
default:true

Whether to scrape links from the page. Links are scraped by default.

​
scrape_images
boolean
default:false

Whether to scrape images from the page. Images are scraped by default.

​
ignored_tags
string[] | null

HTML tags to ignore from the page

​
only_main_content
boolean
default:false

Whether to only scrape the main content of the page. If True, navbars, footers, etc. are excluded.

​
only_images
boolean
default:false

Whether to only scrape images from the page. If True, the page content is excluded.

​
response_format
any | null

The response format to use for the scrape. You can use a Pydantic model or a JSON Schema dict (cf. https://docs.pydantic.dev/latest/concepts/json_schema/#generating-json-schema.)

​
instructions
string | null

Additional instructions to use for the scrape. E.g. 'Extract only the title, date and content of the articles.'

​
use_link_placeholders
boolean
default:false

Whether to use link/image placeholders to reduce the number of tokens in the prompt and hallucinations. However this is an experimental feature and might not work as expected.

Response

Successful Response

​
markdown
string
required

Markdown representation of the extracted data

​
images
ImageData · object[] | null

List of images extracted from the page (ID and download link)

​
structured
StructuredData[BaseModel] · object

Structured data extracted from the page in JSON format