Provide the health status of the service.

The response includes:
- the status of the service
- the number of active sessions
- the current version
- a description of the service

Health Check

HealthResponse

Notte

Quickstart

Python SDK

Authentication

Examples of how to use Notte to navigate a webpage

Navigation

Examples of how to create a scraping flow with Notte

Scraping flow

Learn how Notte manages sessions, billing, and timeouts.

Session Management

This endpoint starts a new session and returns a session ID. An existing session cannot provide a session ID during creation.

Start Session

This endpoint closes an existing session and returns the session ID. A valid session ID must be provided to close the session.

Close Session

This endpoint provides the health status of the service, including the number of active sessions and the current version.

Health

This endpoint observes the actions available on a given URL within the session's environment.

Observe page

This endpoint executes an action in an active browser session and observes the resulting state.

Step in page

This endpoint scrapes data from a specified URL within the session's environment.

Scrape data

Start a new session and returns a session ID. An existing session cannot provide a session ID during creation.

The session request can be parameterized with :
- session_timeout: The session timeout in minutes (reset at every operation done of the session)
- screenshot: Whether to take a screenshot of the browser session.

Close an existing session and returns the session ID.
A valid session ID must be provided to close the session.

Observe the actions available on a given URL within the session's environment.
If a session_id is not provided, a new session is created.

Observe

Execute an action in an active browser session and observe the resulting state.
A valid session ID must be provided to execute an action.

Step

Scrape the content of a given URL within the session's environment.
If a session_id is not provided, a new session is created.
If a session_id is provided and no url is provided, the current page in the session is scraped.

Scrape

Executes actions in the Notte environment based on LLM decisions.

This method demonstrates how to:
1. Parse LLM output into Notte commands
2. Execute those commands in the environment
3. Format the results back into text for the LLM

Users should customize this method to:
- Handle additional Notte endpoints
- Implement custom error handling
- Format observations specifically for their LLM

Args:
    content: The LLM's response containing the desired action

Returns:
    content: Formatted observation from the environment

API Reference

Response