Reference

The Document.

Every acquisition produces this same shape: a scrape, a crawled page, an ingested file. One output contract, so code that handles a scrape handles a crawl result too.

Fields

Field	Type	Detail
url	string	The URL you asked for, canonicalized.
final_url	string	Where it ended up after redirects.
fetched_at	string	ISO 8601, UTC.
status	integer	HTTP status of the final response.
content_type	string	Normalized, without parameters.
depth	integer	Link distance from the seed. Always 0 for a scrape.
title	string	Combined from JSON-LD, OpenGraph, meta, and the document itself.
author	string	Same precedence rules.
published_date	string	Same precedence rules.
language	string	As declared, or detected from the document.
description	string	Meta description or OpenGraph equivalent.
site_name	string	OpenGraph site name where present.
keywords	string[]	As declared by the page.
markdown	string	The primary body representation.
text	string	Plain text, derived from the same extraction.
html	string	Raw HTML. Empty unless include_html was set.
links	Link[]	url, text, internal, nofollow.
images	Image[]	url, alt, title.
tables	Table[]	caption, headers, rows.
records	object[]	Repeated structures, product cards, search results, feed entries.
metadata	object	Anything that does not fit a fixed field, such as json_ld.
trace	Trace	stages[], cache_hit, rendered, redirects[], warnings[].

Precedence

When sources disagree about the title, author or date, the merge is deterministic rather than a judgement call: JSON-LD first, then OpenGraph, then standard meta tags, then the document itself. No model decides what the “real” title is, so the same page always produces the same answer.

Anything without a fixed field lands in metadata, json_ld payloads, feed entry data, and similar, rather than being dropped.

markdown, text, html

markdown is the primary representation and what most pipelines want. text is the same extraction without the formatting. html is the raw body and is empty unless you ask for it. It is by far the largest field, and the key is always present so the shape never changes.

records

Repeated structures found on the page, product cards, search results, feed entries. A listing page yields records; an article usually yields none. This is the field to check when you want the page's data rather than its prose.

A full response

Real output, captured against the project's own test fixture:

200 response

{
  "url": "https://example.com/article",
  "final_url": "https://example.com/article",
  "fetched_at": "2026-07-30T11:43:46.878250Z",
  "status": 200,
  "content_type": "text/html",
  "depth": 0,
  "title": "The Quiet Rise of Underwater Farming",
  "author": "Nadia Okafor",
  "published_date": "2025-11-03",
  "language": "en",
  "description": "How submerged greenhouses are reshaping coastal agriculture.",
  "keywords": ["aquaculture", "farming", "ocean"],
  "site_name": "Ocean Journal",
  "markdown": "# The Quiet Rise of Underwater Farming\n\nTwenty meters beneath …",
  "text": "The Quiet Rise of Underwater Farming\nTwenty meters beneath the Li …",
  "html": "",
  "links": [
    { "url": "https://example.com/", "text": "Home", "internal": true, "nofollow": false }
  ],
  "images": [
    { "url": "https://example.com/images/biosphere.jpg",
      "alt": "A submerged biosphere with basil plants", "title": "" }
  ],
  "tables": [
    { "caption": "Yield comparison",
      "headers": ["Crop", "Land yield", "Biosphere yield"],
      "rows": [["Basil", "1.0x", "1.4x"], ["Strawberry", "1.0x", "1.1x"]] }
  ],
  "records": [],
  "metadata": {
    "json_ld": [
      { "@context": "https://schema.org", "@type": "NewsArticle",
        "headline": "The Quiet Rise of Underwater Farming",
        "author": { "@type": "Person", "name": "Nadia Okafor" },
        "datePublished": "2025-11-03", "inLanguage": "en" }
    ]
  },
  "trace": {
    "stages": [
      { "name": "fetch",    "elapsed_ms": 22.2 },
      { "name": "parse",    "elapsed_ms": 1.3 },
      { "name": "content",  "elapsed_ms": 185.8 },
      { "name": "metadata", "elapsed_ms": 10.0 },
      { "name": "links",    "elapsed_ms": 0.1 },
      { "name": "media",    "elapsed_ms": 0.0 },
      { "name": "records",  "elapsed_ms": 0.1 },
      { "name": "combine",  "elapsed_ms": 0.0 }
    ],
    "cache_hit": false,
    "rendered": false,
    "redirects": [],
    "warnings": []
  }
}

The trace at the end gets its own page.