Page serialization

serialize_sleepwalker_page_content returns Sleepwalker’s normalized page content view for a URL.

Use it when the agent needs to inspect what is on a page before deciding what to do next.

Tool

serialize_sleepwalker_page_content

Required scope:

pages:content:serialize

When To Use It

Use this tool when the user asks for:

• page content extraction
• visible headings and body content
• normalized copy for agent analysis
• a lightweight page view without scoring

Do not use it when the user wants Content Intelligence scoring, AI Visibility prompt suggestions, or a saved run. Use the narrower action that matches the user request.

Inputs

Input	Type	Required	Notes
url	string	Yes	Page URL to serialize.
extraction_mode	string	No	Optional extraction profile. Omit unless instructed otherwise.
max_chars	integer	No	1 to 25,000. Defaults to a bounded chunk.
offset	integer	No	Use for pagination through long content.

Output Shape

The response contains a serialization object:

{
  "serialization": {
    "url": "https://www.sleepwalker.ai",
    "http_status": 200,
    "content_view": "## Page: ...",
    "content_view_offset": 0,
    "content_view_chars": 9025,
    "content_view_truncated": true,
    "next_offset": 1200,
    "billing": {
      "billable": true,
      "credit_action": "pages.content.serialize",
      "estimated_credits": "1.00",
      "credits_enforced": true,
      "status": "settled"
    }
  }
}

Credits

Page content serialization costs 1 credit when it completes successfully.

If the page is blocked, unavailable, or rejected before work begins, credits are not settled for successful work.

See MCP billing for the billing fields.

Pagination

If content_view_truncated is true, call the tool again with offset: next_offset.

Example:

Fetch the next page serialization chunk using next_offset.

Agents should not request the largest possible chunk unless the user needs it.

Example Prompt

Serialize https://www.sleepwalker.ai and summarize the page structure.

Common Errors

Error	Meaning
Missing url	The tool requires a URL.
Invalid extraction_mode	Omit the field or use a supported profile.
Insufficient credits	The account needs more credits for billable actions.
Page blocked	The target site blocked or denied extraction.