AgentBay:Browser Use

Tool

Name

Description

Support Notes

System

create_sandbox

Creates a new AgentBay sandbox and returns its ID.

Supported

get_sandbox_url

Gets the Wuying MCP runtime URL. This URL provides access to the runtime. Each URL is valid for a single use and expires immediately.

Supported

system_screenshot

Captures a full-screen screenshot of the current screen and returns a shareable URL. The screenshot is automatically processed and securely stored. For security, the generated URL expires after 64 minutes.

Supported

kill_sandbox

Releases resources after the task is complete.

Supported

shell

Executes a shell command on the Linux platform with a configurable timeout. After execution, it returns the command's output or an error message.

Supported

Playwright

browser_close

Closes the page.

Supported

browser_resize

Resizes the browser window.

Supported

browser_console_messages

Returns all console messages.

Supported

browser_handle_dialog

Handles a dialog box.

Supported

browser_file_upload

Uploads one or more files.

Supported

browser_install

Installs the browser specified in the configuration. Call this tool if you receive an error that the browser is not installed.

Supported

browser_press_key

Presses a key on the keyboard.

Supported

browser_navigate

Navigates to a URL.

Supported

browser_navigate_back

Goes back to the previous page.

Supported

browser_navigate_forward

Goes forward to the next page.

Supported

browser_network_requests

Returns all network requests since the page was loaded.

Supported

browser_pdf_save

Saves the page as a PDF.

Supported

browser_take_screenshot

Takes a screenshot of the current page. Operations cannot be performed based on the screenshot. To perform operations, use browser_snapshot.

Supported

browser_snapshot

Captures an accessibility snapshot of the current page.

Supported

browser_click

Performs a click operation on the web page.

Supported

browser_drag

Performs a drag-and-drop operation between two elements.

Supported

browser_hover

Hovers the mouse over a page element.

Supported

browser_type

Enters text into an editable element.

Supported

browser_select_option

Selects an option in a drop-down menu.

Supported

browser_tab_list

Lists the browser tabs.

Supported

browser_tab_new

Opens a new tab.

Supported

browser_tab_select

Selects a tab by its index.

Supported

browser_tab_close

Closes a tab.

Supported

browser_generate_playwright_test

Generates a Playwright test for a specified scenario.

Supported

browser_wait_for

Waits for text to appear or disappear, or for a specified time to pass.

Supported

Shell

shell

Executes a shell command and returns the output or an error.

Supported

File system

create_directory

Creates a new directory or ensures a directory exists. Multiple nested directories can be created in one operation. If the directory already exists, this operation succeeds silently. This is ideal for setting up directory structures for projects or ensuring required paths exist. This tool only works in allowed directories.

Supported

edit_file

Performs line-based edits on a text file. Each edit replaces an exact sequence of lines with new content. It returns a git-style diff to show the changes. This tool only works in allowed directories.

Support

get_file_info

Retrieves detailed metadata for a file or directory. It returns comprehensive information, including size, creation time, last modified time, permissions, and type. This tool is ideal for understanding file characteristics without reading the content. This tool only works in allowed directories.

Supported

read_file

Reads the content of a file from the file system. Specify an optional offset (in bytes) to start reading from a specific position, and an optional length (in bytes) to limit the number of bytes to read. If length is omitted or is 0, the file is read to the end. The tool handles various text encodings and provides detailed error messages if a file cannot be read. This tool only works in allowed directories.

Supported

read_multiple_files

Reads the content of multiple files at the same time. This is more efficient than reading files one by one for analysis or comparison. The content of each file is returned with its path as a reference. A failure to read a single file does not stop the entire operation. This tool only works in allowed directories.

Supported

list_directory

Gets a detailed list of all files and directories in a specified path. The results clearly distinguish between files and directories with [FILE] and [DIR] prefixes. This tool is useful for understanding directory structures and finding specific files. This tool only works in allowed directories.

Supported

move_file

Moves or renames files and directories. A file can be moved to a different directory and renamed in one operation. If the destination already exists, the operation fails. The tool works between different directories and can be used for simple renames within the same directory. Both the source and destination must be in allowed directories.

Supported

search_files

Recursively searches for files and directories that match a pattern. It searches all subdirectories from a starting path. The search is case-sensitive and matches partial names. It returns the full paths of all matching items. This is ideal for finding files when you do not know their exact location. This tool only searches in allowed directories.

Supported

write_file

Creates a new file or writes to an existing file. Choose to either completely overwrite the file or append to its end by specifying the mode parameter. Use the overwrite mode (default) to clear the file before writing, or use the append mode to add content to the end of the file. The tool handles text content with the proper encoding. This tool only works in allowed directories.

Supported

PageUseAgent

page_use_act

Executes one or more operations on the current web page and blocks until all operations are complete.

This method immediately executes the provided operations through the current agent and returns the final result after the entire sequence is complete. It can handle single interactions along with a series of continuous operations.

Supported

page_use_screenshot

Captures a screenshot of the current web page.

This tool provides flexible screenshot features, including full-page screenshots, cropping to a specified area, and image quality settings. The result is returned as a data URL (such as data:image/png;base64), which can be used directly in frontend applications.

Supported

page_use_close_session

Closes the current browser proxy session.

This terminates the browser process managed by the proxy and releases related resources.

Supported

page_use_observe

Identifies and locates interactive user interface elements.

This tool helps you discover and describe elements that can be interacted with (such as buttons or input boxes) for use in subsequent operations.

To find operable elements, use the observe tool.

To extract structured data or text content, use the extract tool instead.

Supported

page_use_extract_async

Asynchronously extracts information from a web page based on the provided instructions. It returns a structured result in the specified schema format.

Supported

page_use_act_async

Asynchronously starts one or more operations on the current web page and returns a task_id for progress polling.

Unlike page_use_act, this method does not block until the operations are complete. It starts execution immediately, returns a unique task_id, and lets the client poll for intermediate progress and the final result using page_use_get_act_result.

Supported

page_use_get_act_result

Gets the current progress or final result of an asynchronous operation task.

Use this method with the task_id obtained from a previous call to page_use_act_async. It returns the current status of the task. If the execution is complete (is_done), it includes the result content.

Supported

page_use_get_extract_result

An asynchronous function that gets the result of an extract task.

Supported

page_use_navigate

Navigates to the specified URL in the browser.

Supported

Name

Description

Parameters

create_sandbox

Creates a new AgentBay sandbox and returns its ID.

{
  "type": "object",
  "properties": {},
  "required": []
}

get_sandbox_url

Gets the Wuying MCP runtime URL. This URL provides access to the runtime. Each URL is valid for a single use and expires immediately.

{
  "type": "object",
  "properties": {
    "sandbox_id": {
      "type": "string",
      "title": "Sandbox Id",
      "description": "The sandbox ID is the identifier for the tool execution environment. This sandbox_id comes from the create_sandbox tool."
    }
  },
  "required": [
    "sandbox_id"
  ]
}

system_screenshot

Captures a full-screen screenshot of the current screen and returns a shareable URL. The screenshot is automatically processed and securely stored. For security, the generated URL expires after 64 minutes.

{
  "type": "object",
  "properties": {
    "sandbox_id": {
      "type": "string",
      "title": "Sandbox Id",
      "description": "The sandbox ID is the identifier for the tool execution environment. This sandbox_id comes from the create_sandbox tool."
    }
  },
  "required": [
    "sandbox_id"
  ]
}

kill_sandbox

Releases resources after the task is complete.

{
  "type": "object",
  "properties": {
    "sandbox_id": {
      "type": "string",
      "title": "Sandbox Id",
      "description": "The sandbox ID is the identifier for the tool execution environment. This sandbox_id comes from the create_sandbox tool."
    }
  },
  "required": [
    "sandbox_id"
  ]
}

shell

Executes a shell command on the Linux platform with a configurable timeout. After execution, it returns the command's output or an error message.

{
  "type": "object",
  "required": [
    "sandbox_id",
    "command",
    "timeout_ms"
  ],
  "properties": {
    "sandbox_id": {
      "type": "string",
      "title": "Sandbox Id",
      "description": "The sandbox ID is the identifier for the tool execution environment. This sandbox_id comes from the create_sandbox tool."
    },
    "timeout_ms": {
      "default": 1000,
      "description": "The command execution timeout in milliseconds. If not specified, a default value is used, such as 1000 ms.",
      "type": "integer"
    },
    "command": {
      "description": "The command entered by the client.",
      "type": "string"
    }
  }
}

Name

Description

Parameters

browser_close

Closes the page

{
  "inputSchema": {
    "type": "object",
    "properties": {},
    "additionalProperties": false,
    "$schema": "http://json-schema.org/draft-07/schema#"
  }
}

browser_resize

Resizes the browser window

{
  "inputSchema": {
    "type": "object",
    "properties": {
      "width": {
        "type": "number",
        "description": "The width of the browser window."
      },
      "height": {
        "type": "number",
        "description": "The height of the browser window."
      }
    },
    "required": ["width", "height"],
    "additionalProperties": false,
    "$schema": "http://json-schema.org/draft-07/schema#"
  }
}

browser_console_messages

Returns all console messages

{
  "inputSchema": {
    "type": "object",
    "properties": {},
    "additionalProperties": false,
    "$schema": "http://json-schema.org/draft-07/schema#"
  }
}

browser_handle_dialog

Handles a dialog box

{
  "inputSchema": {
    "type": "object",
    "properties": {
      "accept": {
        "type": "boolean",
        "description": "Specifies whether to accept the dialog box."
      },
      "promptText": {
        "type": "string",
        "description": "The prompt text in a prompt dialog box."
      }
    },
    "required": ["accept"],
    "additionalProperties": false,
    "$schema": "http://json-schema.org/draft-07/schema#"
  }
}

browser_file_upload

Uploads one or more files

{
  "inputSchema": {
    "type": "object",
    "properties": {
      "paths": {
        "type": "array",
        "items": {
          "type": "string"
        },
        "description": "The absolute paths of the files to upload. You can specify a single file or multiple files."
      }
    },
    "required": ["paths"],
    "additionalProperties": false,
    "$schema": "http://json-schema.org/draft-07/schema#"
  }
}

browser_install

Installs the browser specified in the configuration. Call this tool if you receive an error that the browser is not installed.

{
  "inputSchema": {
    "type": "object",
    "properties": {},
    "additionalProperties": false,
    "$schema": "http://json-schema.org/draft-07/schema#"
  }
}

browser_press_key

Presses a key on the keyboard

{
  "inputSchema": {
    "type": "object",
    "properties": {
      "key": {
        "type": "string",
        "description": "The name of the key to press or the character to generate, such as `ArrowLeft` or `a`."
      }
    },
    "required": ["key"],
    "additionalProperties": false,
    "$schema": "http://json-schema.org/draft-07/schema#"
  }
}

browser_navigate

Navigates to a URL

{
  "inputSchema": {
    "type": "object",
    "properties": {
      "url": {
        "type": "string",
        "description": "The URL to navigate to."
      }
    },
    "required": ["url"],
    "additionalProperties": false,
    "$schema": "http://json-schema.org/draft-07/schema#"
  }
}

browser_navigate_back

Goes back to the previous page

{
  "inputSchema": {
    "type": "object",
    "properties": {},
    "additionalProperties": false,
    "$schema": "http://json-schema.org/draft-07/schema#"
  }
}

browser_navigate_forward

Goes forward to the next page

{
  "inputSchema": {
    "type": "object",
    "properties": {},
    "additionalProperties": false,
    "$schema": "http://json-schema.org/draft-07/schema#"
  }
}

browser_network_requests

Returns all network requests since the page was loaded

{
  "inputSchema": {
    "type": "object",
    "properties": {},
    "additionalProperties": false,
    "$schema": "http://json-schema.org/draft-07/schema#"
  }
}

browser_pdf_save

Saves the page as a PDF

{
  "inputSchema": {
    "type": "object",
    "properties": {
      "filename": {
        "type": "string",
        "description": "The file name to save the PDF to. If not specified, the default is `page-{timestamp}.pdf`."
      }
    },
    "additionalProperties": false,
    "$schema": "http://json-schema.org/draft-07/schema#"
  }
}

browser_take_screenshot

Takes a screenshot of the current page. Operations cannot be performed based on the screenshot. Use browser_snapshot for operations.

{
  "inputSchema": {
    "type": "object",
    "properties": {
      "raw": {
        "type": "boolean",
        "description": "Specifies whether to return uncompressed data in PNG format. The default value is false, which returns a JPEG image."
      },
      "filename": {
        "type": "string",
        "description": "The file name to save the screenshot to. If not specified, the default is `page-{timestamp}.{png|jpeg}`."
      },
      "element": {
        "type": "string",
        "description": "A human-readable description of the element to get permission to take its screenshot. If not provided, the viewport is captured. If element is provided, ref must also be provided."
      },
      "ref": {
        "type": "string",
        "description": "A reference to the target element in the page snapshot. If not provided, the viewport is captured. If ref is provided, element must also be provided."
      }
    },
    "additionalProperties": false,
    "$schema": "http://json-schema.org/draft-07/schema#"
  }
}

browser_snapshot

Captures an accessibility snapshot of the current page.

{
  "inputSchema": {
    "type": "object",
    "properties": {},
    "additionalProperties": false,
    "$schema": "http://json-schema.org/draft-07/schema#"
  }
}

browser_click

Performs a click operation on the web page

{
  "inputSchema": {
    "type": "object",
    "properties": {
      "element": {
        "type": "string",
        "description": "A human-readable description of the element to get permission to interact with it."
      },
      "ref": {
        "type": "string",
        "description": "A reference to the target element in the page snapshot."
      }
    },
    "required": ["element", "ref"],
    "additionalProperties": false,
    "$schema": "http://json-schema.org/draft-07/schema#"
  }
}

browser_drag

Performs a drag-and-drop operation between two elements

{
  "inputSchema": {
    "type": "object",
    "properties": {
      "startElement": {
        "type": "string",
        "description": "A human-readable description of the source element to get permission to interact with it."
      },
      "startRef": {
        "type": "string",
        "description": "A reference to the source element in the page snapshot."
      },
      "endElement": {
        "type": "string",
        "description": "A human-readable description of the target element to get permission to interact with it."
      },
      "endRef": {
        "type": "string",
        "description": "A reference to the target element in the page snapshot."
      }
    },
    "required": ["startElement", "startRef", "endElement", "endRef"],
    "additionalProperties": false,
    "$schema": "http://json-schema.org/draft-07/schema#"
  }
}

browser_hover

Hovers the mouse over a page element

{
  "inputSchema": {
    "type": "object",
    "properties": {
      "element": {
        "type": "string",
        "description": "A human-readable description of the element to get permission to interact with it."
      },
      "ref": {
        "type": "string",
        "description": "A reference to the target element in the page snapshot."
      }
    },
    "required": ["element", "ref"],
    "additionalProperties": false,
    "$schema": "http://json-schema.org/draft-07/schema#"
  }
}

browser_type

Enters text into an editable element

{
  "inputSchema": {
    "type": "object",
    "properties": {
      "element": {
        "type": "string",
        "description": "A human-readable description of the element to get permission to interact with it."
      },
      "ref": {
        "type": "string",
        "description": "A reference to the target element in the page snapshot."
      },
      "text": {
        "type": "string",
        "description": "The text to enter into the element."
      },
      "submit": {
        "type": "boolean",
        "description": "Specifies whether to submit the entered text by pressing Enter."
      },
      "slowly": {
        "type": "boolean",
        "description": "Specifies whether to enter text one character at a time. This is useful for triggering key handlers on the page. By default, the entire text is filled at once."
      }
    },
    "required": ["element", "ref", "text"],
    "additionalProperties": false,
    "$schema": "http://json-schema.org/draft-07/schema#"
  }
}

browser_select_option

Selects an option in a drop-down menu

{
  "inputSchema": {
    "type": "object",
    "properties": {
      "element": {
        "type": "string",
        "description": "A human-readable description of the element to get permission to interact with it."
      },
      "ref": {
        "type": "string",
        "description": "A reference to the target element in the page snapshot."
      },
      "values": {
        "type": "array",
        "items": {
          "type": "string"
        },
        "description": "An array of values to select in the drop-down menu. You can specify a single value or multiple values."
      }
    },
    "required": ["element", "ref", "values"],
    "additionalProperties": false,
    "$schema": "http://json-schema.org/draft-07/schema#"
  }
}

browser_tab_list

Lists the browser tabs

{
  "inputSchema": {
    "type": "object",
    "properties": {},
    "additionalProperties": false,
    "$schema": "http://json-schema.org/draft-07/schema#"
  }
}

browser_tab_new

Opens a new tab

{
  "inputSchema": {
    "type": "object",
    "properties": {
      "url": {
        "type": "string",
        "description": "The URL to navigate to in the new tab. If not provided, the new tab is blank."
      }
    },
    "additionalProperties": false,
    "$schema": "http://json-schema.org/draft-07/schema#"
  }
}

browser_tab_select

Selects a tab by its index

{
  "inputSchema": {
    "type": "object",
    "properties": {
      "index": {
        "type": "number",
        "description": "The index of the tab to select."
      }
    },
    "required": ["index"],
    "additionalProperties": false,
    "$schema": "http://json-schema.org/draft-07/schema#"
  }
}

browser_tab_close

Closes a tab

{
  "inputSchema": {
    "type": "object",
    "properties": {
      "index": {
        "type": "number",
        "description": "The index of the tab to close. If not provided, the current tab is closed."
      }
    },
    "additionalProperties": false,
    "$schema": "http://json-schema.org/draft-07/schema#"
  }
}

browser_generate_playwright_test

Generates a Playwright test for a specified scenario

{
  "inputSchema": {
    "type": "object",
    "properties": {
      "name": {
        "type": "string",
        "description": "The name of the test."
      },
      "description": {
        "type": "string",
        "description": "The description of the test."
      },
      "steps": {
        "type": "array",
        "items": {
          "type": "string"
        },
        "description": "The steps of the test."
      }
    },
    "required": ["name", "description", "steps"],
    "additionalProperties": false,
    "$schema": "http://json-schema.org/draft-07/schema#"
  }
}

browser_wait_for

Waits for text to appear or disappear, or for a specified time to pass

{
  "inputSchema": {
    "type": "object",
    "properties": {
      "time": {
        "type": "number",
        "description": "The time to wait in seconds."
      },
      "text": {
        "type": "string",
        "description": "The text to wait for."
      },
      "textGone": {
        "type": "string",
        "description": "The text to wait for to disappear."
      }
    },
    "additionalProperties": false,
    "$schema": "http://json-schema.org/draft-07/schema#"
  }
}

Name

Description

Parameters

shell

Executes a shell command and returns the output or an error.

{
  "inputSchema": {
    "properties": {
      "command": {
        "description": "The command entered by the client.",
        "type": "string"
      },
      "timeout_ms": {
        "default": 1000,
        "description": "The command execution timeout in milliseconds. If not specified, a default value is used, such as 1000 ms.",
        "type": "integer"
      }
    },
    "required": ["command", "timeout_ms"],
    "type": "object"
  }
}

Name

Description

Parameters

create_directory

Creates a new directory or ensures a directory exists. Multiple nested directories can be created in one operation. If the directory already exists, this operation succeeds silently. This is ideal for setting up directory structures for projects or ensuring required paths exist. This tool only works in allowed directories.

{
  "inputSchema": {
    "properties": {
      "path": {
        "description": "The path of the directory to create.",
        "type": "string"
      }
    },
    "required": ["path"],
    "type": "object"
  }
}

edit_file

Performs line-based edits on a text file. Each edit replaces an exact sequence of lines with new content. It returns a git-style diff to show the changes. This tool only works in allowed directories.

{
  "inputSchema": {
    "properties": {
      "dryRun": {
        "default": false,
        "description": "Previews changes in a git-style diff format.",
        "type": "boolean"
      },
      "edits": {
        "items": {
          "properties": {
            "newText": {
              "description": "The new text to replace the old text.",
              "type": "string"
            },
            "oldText": {
              "description": "The text to search for. It must be an exact match.",
              "type": "string"
            }
          },
          "required": ["oldText", "newText"],
          "type": "object"
        },
        "type": "array"
      },
      "path": {
        "description": "The path of the file to edit.",
        "type": "string"
      }
    },
    "required": ["path", "edits"],
    "type": "object"
  }
}

get_file_info

Retrieves detailed metadata for a file or directory. It returns comprehensive information, including size, creation time, last modified time, permissions, and type. This tool is ideal for understanding file characteristics without reading the content. This tool only works in allowed directories.

{
  "inputSchema": {
    "properties": {
      "path": {
        "description": "The path of the file or directory to inspect.",
        "type": "string"
      }
    },
    "required": ["path"],
    "type": "object"
  }
}

read_file

Reads the content of a file from the file system. You can specify an optional 'offset' (in bytes) to start reading from a specific position, and an optional 'length' (in bytes) to limit how many bytes to read. If 'length' is omitted or 0, the file will be read to the end. Handles various text encodings and provides detailed error messages if the file cannot be read. Only works within allowed directories.

{
  "inputSchema": {
    "properties": {
      "length": {
        "description": "The number of bytes to read. If omitted or set to 0, the file is read to the end.",
        "minimum": 0,
        "type": "integer"
      },
      "offset": {
        "default": 0,
        "description": "The byte offset from which to start reading.",
        "minimum": 0,
        "type": "integer"
      },
      "path": {
        "description": "The path of the file to read.",
        "type": "string"
      }
    },
    "required": ["path"],
    "type": "object"
  }
}

read_multiple_files

Reads the content of multiple files at the same time. This is more efficient than reading files one by one for analysis or comparison. The content of each file is returned with its path as a reference. A failure to read a single file does not stop the entire operation. This tool only works in allowed directories.

{
  "inputSchema": {
    "properties": {
      "paths": {
        "description": "An array of file paths to read.",
        "items": {
          "type": "string"
        },
        "type": "array"
      }
    },
    "required": ["paths"],
    "type": "object"
  }
}

list_directory

Gets a detailed list of all files and directories in a specified path. The results clearly distinguish between files and directories with [FILE] and [DIR] prefixes. This tool is useful for understanding directory structures and finding specific files. This tool only works in allowed directories.

{
  "inputSchema": {
    "properties": {
      "path": {
        "description": "The path of the directory to list.",
        "type": "string"
      }
    },
    "required": ["path"],
    "type": "object"
  }
}

move_file

Moves or renames files and directories. A file can be moved to a different directory and renamed in one operation. If the destination already exists, the operation fails. The tool works between different directories and can be used for simple renames within the same directory. Both the source and destination must be in allowed directories.

{
  "inputSchema": {
    "properties": {
      "destination": {
        "description": "The destination file or directory path.",
        "type": "string"
      },
      "source": {
        "description": "The source file or directory path.",
        "type": "string"
      }
    },
    "required": ["source", "destination"],
    "type": "object"
  }
}

search_files

Recursively searches for files and directories that match a pattern. It searches all subdirectories from a starting path. The search is case-sensitive and matches partial names. It returns the full paths of all matching items. This is ideal for finding files when you do not know their exact location. This tool only searches in allowed directories.

{
  "inputSchema": {
    "properties": {
      "excludePatterns": {
        "default": [],
        "description": "The patterns to exclude (optional).",
        "items": {
          "type": "string"
        },
        "type": "array"
      },
      "path": {
        "description": "The directory path from which to start the search.",
        "type": "string"
      },
      "pattern": {
        "description": "The pattern to match.",
        "type": "string"
      }
    },
    "required": ["path", "pattern"],
    "type": "object"
  }
}

write_file

Creates a new file or writes to an existing file. Choose to either completely overwrite the file or append to its end by specifying the mode parameter. Use the overwrite mode (default) to clear the file before writing, or use the append mode to add content to the end of the file. The tool handles text content with the proper encoding. This tool only works in allowed directories.

{
  "inputSchema": {
    "properties": {
      "content": {
        "description": "The content to write.",
        "type": "string"
      },
      "mode": {
        "default": "overwrite",
        "description": "The write mode. 'overwrite' clears the file, and 'append' adds content to the end of the file.",
        "enum": ["overwrite", "append"],
        "type": "string"
      },
      "path": {
        "description": "The path of the file to write to.",
        "type": "string"
      }
    },
    "required": ["path", "content"],
    "type": "object"
  }
}

Name

Description

Parameters

page_use_act

Executes one or more operations on the current web page and blocks until all operations are complete.

This method immediately executes the provided operations through the current agent and returns the final result after the entire sequence is complete. It can handle single interactions along with a series of continuous operations.

{
  "type": "object",
  "title": "page_use_actArguments",
  "required": [
    "sandbox_id",
    "action"
  ],
  "properties": {
    "sandbox_id": {
      "type": "string",
      "title": "Sandbox Id",
      "description": "The sandbox ID is the identifier for the tool execution environment. This sandbox_id comes from the create_sandbox tool."
    },
    "dom_settle_timeout_ms": {
      "type": "integer",
      "title": "Dom Settle Timeout Ms"
    },
    "page_id": {
      "anyOf": [
        {
          "type": "string"
        },
        {
          "type": "null"
        }
      ],
      "title": "Page Id"
    },
    "variables": {
      "additionalProperties": {
        "type": "string"
      },
      "type": "object",
      "title": "Variables"
    },
    "action": {
      "type": "string",
      "title": "Action"
    },
    "context_id": {
      "default": 0,
      "type": "integer",
      "title": "Context Id"
    },
    "iframe": {
      "type": "boolean",
      "title": "Iframe"
    },
    "timeout_ms": {
      "type": "integer",
      "title": "Timeout Ms"
    },
    "use_vision": {
      "default": false,
      "type": "boolean",
      "title": "Use Vision"
    }
  }
}

page_use_screenshot

Captures a screenshot of the current web page.

This tool provides flexible screenshot features, including full-page screenshots, cropping to a specified area, and image quality settings. The result is returned as a data URL (such as data:image/png;base64), which can be used directly in frontend applications.

{
  "type": "object",
  "title": "page_use_screenshotArguments",
  "properties": {
    "sandbox_id": {
      "type": "string",
      "title": "Sandbox Id",
      "description": "The sandbox ID is the identifier for the tool execution environment. This sandbox_id comes from the create_sandbox tool."
    },
    "page_id": {
      "anyOf": [
        {
          "type": "string"
        },
        {
          "type": "null"
        }
      ],
      "title": "Page Id"
    },
    "full_page": {
      "default": true,
      "type": "boolean",
      "title": "Full Page"
    },
    "context_id": {
      "default": 0,
      "type": "integer",
      "title": "Context Id"
    },
    "clip": {
      "additionalProperties": {
        "type": "number"
      },
      "type": "object",
      "title": "Clip"
    },
    "timeout": {
      "type": "integer",
      "title": "Timeout"
    },
    "quality": {
      "default": 80,
      "type": "integer",
      "title": "Quality"
    }
  },
  "required": [
    "sandbox_id"
  ]
}

page_use_close_session

Closes the current browser proxy session.

This terminates the browser process managed by the proxy and releases related resources.

{
  "type": "object",
  "title": "page_use_close_sessionArguments",
  "properties": {
    "sandbox_id": {
      "type": "string",
      "title": "Sandbox Id",
      "description": "The sandbox ID is the identifier for the tool execution environment. This sandbox_id comes from the create_sandbox tool."
    }
  },
  "required": [
    "sandbox_id"
  ]
}

page_use_observe

Identifies and locates interactive user interface elements.

This tool helps you discover and describe elements that can be interacted with (such as buttons or input boxes) for use in subsequent operations.

To find operable elements, use the observe tool.

To extract structured data or text content, use the extract tool instead.

{
  "type": "object",
  "title": "page_use_observeArguments",
  "required": [
    "sandbox_id",
    "instruction"
  ],
  "properties": {
    "sandbox_id": {
      "type": "string",
      "title": "Sandbox Id",
      "description": "The sandbox ID is the identifier for the tool execution environment. This sandbox_id comes from the create_sandbox tool."
    },
    "dom_settle_timeout_ms": {
      "type": "integer",
      "title": "Dom Settle Timeout Ms"
    },
    "page_id": {
      "anyOf": [
        {
          "type": "string"
        },
        {
          "type": "null"
        }
      ],
      "title": "Page Id"
    },
    "instruction": {
      "type": "string",
      "title": "Instruction"
    },
    "context_id": {
      "default": 0,
      "type": "integer",
      "title": "Context Id"
    },
    "iframe": {
      "default": false,
      "type": "boolean",
      "title": "Iframe"
    },
    "use_vision": {
      "default": false,
      "type": "boolean",
      "title": "Use Vision"
    }
  }
}

page_use_extract_async

Asynchronously extracts information from a web page based on the provided instructions. It returns a structured result in the specified schema format.

{
  "type": "object",
  "title": "page_use_extract_asyncArguments",
  "required": [
    "sandbox_id",
    "instruction",
    "field_schema"
  ],
  "properties": {
    "sandbox_id": {
      "type": "string",
      "title": "Sandbox Id",
      "description": "The sandbox ID is the identifier for the tool execution environment. This sandbox_id comes from the create_sandbox tool."
    },
    "dom_settle_timeout_ms": {
      "type": "integer",
      "title": "Dom Settle Timeout Ms"
    },
    "page_id": {
      "anyOf": [
        {
          "type": "string"
        },
        {
          "type": "null"
        }
      ],
      "title": "Page Id"
    },
    "use_text_extract": {
      "default": false,
      "type": "boolean",
      "title": "Use Text Extract"
    },
    "field_schema": {
      "type": "string",
      "title": "Field Schema"
    },
    "instruction": {
      "type": "string",
      "title": "Instruction"
    },
    "selector": {
      "type": "string",
      "title": "Selector"
    },
    "context_id": {
      "default": 0,
      "type": "integer",
      "title": "Context Id"
    },
    "iframe": {
      "default": false,
      "type": "boolean",
      "title": "Iframe"
    },
    "use_vision": {
      "default": false,
      "type": "boolean",
      "title": "Use Vision"
    }
  }
}

page_use_act_async

Asynchronously starts one or more operations on the current web page and returns a task_id for progress polling.

Unlike page_use_act, this method does not block until the operations are complete. It starts execution immediately, returns a unique task_id, and lets the client poll for intermediate progress and the final result using page_use_get_act_result.

{
  "type": "object",
  "title": "page_use_act_asyncArguments",
  "required": [
    "sandbox_id",
    "action"
  ],
  "properties": {
    "sandbox_id": {
      "type": "string",
      "title": "Sandbox Id",
      "description": "The sandbox ID is the identifier for the tool execution environment. This sandbox_id comes from the create_sandbox tool."
    },
    "dom_settle_timeout_ms": {
      "type": "integer",
      "title": "Dom Settle Timeout Ms"
    },
    "page_id": {
      "anyOf": [
        {
          "type": "string"
        },
        {
          "type": "null"
        }
      ],
      "title": "Page Id"
    },
    "variables": {
      "additionalProperties": {
        "type": "string"
      },
      "type": "object",
      "title": "Variables"
    },
    "action": {
      "type": "string",
      "title": "Action"
    },
    "context_id": {
      "default": 0,
      "type": "integer",
      "title": "Context Id"
    },
    "iframe": {
      "type": "boolean",
      "title": "Iframe"
    },
    "timeout_ms": {
      "type": "integer",
      "title": "Timeout Ms"
    },
    "use_vision": {
      "default": false,
      "type": "boolean",
      "title": "Use Vision"
    }
  }
}

page_use_get_act_result

Gets the current progress or final result of an asynchronous operation task.

Use this method with the task_id obtained from a previous call to page_use_act_async. It returns the current status of the task. If the execution is complete (is_done), it includes the result content.

{
  "type": "object",
  "title": "page_use_get_act_resultArguments",
  "required": [
    "sandbox_id",
    "task_id"
  ],
  "properties": {
    "sandbox_id": {
      "type": "string",
      "title": "Sandbox Id",
      "description": "The sandbox ID is the identifier for the tool execution environment. This sandbox_id comes from the create_sandbox tool."
    },
    "task_id": {
      "type": "string",
      "title": "Task Id"
    }
  }
}

page_use_get_extract_result

An asynchronous function that gets the result of an extract task.

{
  "type": "object",
  "title": "page_use_get_extract_resultArguments",
  "required": [
    "sandbox_id",
    "task_id"
  ],
  "properties": {
    "sandbox_id": {
      "type": "string",
      "title": "Sandbox Id",
      "description": "The sandbox ID is the identifier for the tool execution environment. This sandbox_id comes from the create_sandbox tool."
    },
    "task_id": {
      "type": "string",
      "title": "Task Id"
    }
  }
}

page_use_get_extract_result

Navigates to the specified URL in the browser.

{
  "type": "object",
  "title": "page_use_navigateArguments",
  "required": [
    "sandbox_id",
    "url"
  ],
  "properties": {
    "sandbox_id": {
      "type": "string",
      "title": "Sandbox Id",
      "description": "The sandbox ID is the identifier for the tool execution environment. This sandbox_id comes from the create_sandbox tool."
    },
    "url": {
      "type": "string",
      "title": "Url"
    }
  }
}