Structured Contents Initiative

More Structured Data, less BLOBs

Wikipedia uses wikitext, a markup language designed for formatting page content. While it has proven useful for editors authoring wiki articles, it creates complexity for developers parsing articles at scale.

When Wikimedia Enterprise launched in 2021, we built it to serve high-volume, high-frequency users of Wikimedia data. As part of that effort we also improved parsing by providing HTML blobs, a format that developers are more familiar with and for which many parsing libraries already exist.

The Structured Contents Initiative is the next step in serving easy-to-parse Wikimedia data. Currently in beta, it extracts infoboxes, sections, tables, references, and more from raw wikitext and HTML and delivers them as structured, machine-readable JSON.

What's on this page:

What's Available Now
Showcase: BLOBs vs Structured
How to access Structured Contents
Shaping Structured Contents Together
Payload Example
Keep exploring

What's Available Now

Languages: English, Spanish, Portuguese, French, German, Italian, Indonesian, Dutch, and Welsh.

APIs: On-demand & Snapshot

Structured Contents currently extracts the following article parts into JSON:

abstract
description
main image
infobox
sections
images
lists
citations & references
tables

Below is one article's infobox twice over: the wikitext the payload carries, and the JSON Structured Contents returns for the same rows. Both are real API output, abbreviated only by eliding whole entries.

One side needs a parser to find where the infobox even begins. The other has already named every row in it.

"{{Short description|American-born French entertainer (1906–1975)}}\n{{Other people}}\n{{Use American English|date=January 2026}}\n{{Use mdy dates|date=June 2026}}\n{{Infobox person\n| name               = Josephine Baker\n| image              = File:Baker Harcourt 1940 2.jpg\n| caption            = Baker in 1940\n| birth_name         = Freda Josephine McDonald\n| birth_date         = {{birth date|mf=yes|1906|06|03}}\n| birth_place        = [[St. Louis]], Missouri, U.S.\n| death_date         = {{death date and age|mf=yes|1975|04|12|1906|06|03}}\n| death_place        = Paris, France\n| resting_place      = [[Monaco Cemetery]] \n| citizenship        = {{plainlist|\n* United States (until 1937)<!-- renounced US citizenship in 1937 -->\n* France (from 1937)\n}}\n| occupation         = [[Vedette (cabaret)|Vedette]], singer, dancer, actress, [[civil rights activist]], [[French Resistance]] agent\n| years_active       = 1921–1975\n| spouse             = {{Plain list|\n* {{marriage|Willie Wells|1919|1919|reason=div}}\n* {{marriage|William Baker|1921|1925|reason=div}}\n* {{marriage|Jean Lion|1937|1940|reason=div}}\n* {{marriage|[[Jo Bouillon]]|1947|1961|reason=div}}\n}}\n| partner            = Robert Brady <br />(1973–1975)\n| children           = 12{{efn|[[Jean-Claude Baker]] presented himself as her foster son, however his claim had been contested by Baker’s children.<ref>{{cite news |author=<!--none--> |url=https://www.bienpublic.com/actualite/2015/01/16/jean-claude-bouillon-baker-un-celebre-dijonnais-s-est-eteint |work=Le bien public |title=Jean-Claude Rouzaud s'est éteint, pas Jean-Claude Bouillon-Baker |date=January 22, 2015 |access-date=April 29, 2022 |archive-date=May 28, 2023 |archive-url=https://web.archive.org/web/20230528020932/https://www.bienpublic.com/actualite/2015/01/16/jean-claude-bouillon-baker-un-celebre-dijonnais-s-est-eteint |url-status=live }}</ref><ref>{{cite news |last=Dumas |first=Thierry |url=https://www.sudouest.fr/dordogne/sarlat-la-caneda/je-lui-dois-bien-plus-que-ca-9224950.php |work=Sud ouest |title='Je lui dois bien plus que ça' |date=March 24, 2012 |access-date=April 29, 2022 |archive-date=April 29, 2022 |archive-url=https://web.archive.org/web/20220429152001/https://www.sudouest.fr/dordogne/sarlat-la-caneda/je-lui-dois-bien-plus-que-ca-9224950.php |url-status=live }}</ref>}}\n| signature          = Signature de Joséphine Baker - Archives nationales (France).png\n| module             = {{Infobox musical artist|embed=yes\n | background    = solo_singer\n | genre         = {{hlist|[[Cabaret]]|[[music hall]]|[[French pop]]|[[French jazz]]}}\n | instrument    = Vocals\n | label         = {{hlist|[[Columbia Records|Columbia]]| [[Mercury Records|Mercury]]|[[RCA Records|RCA Victor]]\n}}}}\n}}\n\n'''Freda Josephine Baker''' ({{née|'''McDonald'''}}; June 3, 1906 – April 12, 1975), also spelled '''Joséphine Baker''',<ref>{{cite web |title=French Women & Feminists in History: A Resource Guide. Joséphine Baker |url=https://guides.loc.gov/feminism-french-women-history/famous/josephine-baker |website=Research Guides |publ [...]"

"infoboxes": [
  {
    "type": "infobox",
    "name": "Infobox person",
    "has_parts": [
      {
        "type": "section",
        "name": "Josephine Baker",
        "has_parts": [
          {
            "type": "image",
            "value": "Baker in 1940",
            "images": [...]
          },
          {
            "type": "field",
            "name": "Born",
            "value": "Freda Josephine McDonald June 3, 1906 St. Louis, Missouri, U.S.",
            "links": [...]
          },
          {
            "type": "field",
            "name": "Died",
            "value": "April 12, 1975 (aged 68) Paris, France"
          },
          [...]
        ]
      }
    ]
  }
]

For a full explanation of the Structured Contents response data schema see the structured_content model in our API reference.

Showcase: BLOBs vs Structured Contents

Below are examples drawn from English Wikipedia articles, each feature shown side by side comparing the raw HTML and wikitext BLOBs versus the clean JSON output from Structured Contents. These examples make it clear how the data is transformed and why it is easier for developers to use. Some of the payload output in these examples have been truncated (using [...]).

Article Description

Short description of what the article is about, more concise than abstract

[...]><div class=\"shortdescription nomobile noexcerpt noprint searchaux\" style=\"display:none\" about=\"#mwt1\" typeof=\"mw:Transclusion\" data-mw='{\"parts\":[{\"template\":{\"target\":{\"wt\":\"short description\",\"href\":\"./Template:Short_description\"},\"params\":{\"1\":{\"wt\":\"American-born French entertainer (1906–1975)\"}},\"i\":0}}]}' id=\"mwAg\">American-born French entertainer (1906–1975)</div[...]

{{short description|American-born French entertainer (1906–1975)}}\n

"description": "American-born French entertainer (1906–1975)"

Infoboxes

Parsed infoboxes of the article, as recursive trees of parts

data-mw-deduplicate=\"TemplateStyles:r1295905060\" typeof=\"mw:Extension/templatestyles mw:Transclusion\" about=\"#mwt6\" data-mw='{\"name\":\"templatestyles\",\"attrs\":{\"src\":\"Module:Infobox/styles.css\"},\"body\":{\"extsrc\":\"\"},\"parts\":[{\"template\":{\"target\":{\"wt\":\"Infobox person\\n\",\"href\":\"./Template:Infobox_person\"},\"params\":{\"name\":{\"wt\":\"Josephine Baker\"},\"image\":{\"wt\":\"File:Baker Harcourt 1940 2.jpg\"},\"caption\":{\"wt\":\"Baker in 1940\"},\"birth_name\":{\"wt\":\"Freda Josephine McDonald\"},\"birth_date\":{\"wt\":\"{{birth date|mf=yes|1906|06|03}}\"},\"birth_place\":{\"wt\":\"[[St. Louis]], Missouri, U.S.\"}
[...]

{{Infobox person\n| name               = Josephine Baker\n| image              = File:Baker Harcourt 1940 2.jpg\n| caption            = Baker in 1940\n| birth_name         = Freda Josephine McDonald\n| birth_date         = {{birth date|mf=yes|1906|06|03}}\n| birth_place        = [[St. Louis]], Missouri, U.S.\n| [...]

"infoboxes": [{
  "name": "Infobox person",
  "type": "infobox",
  "has_parts": [
    {
      "name": "Josephine Baker",
      "type": "section",
      "has_parts": [
        {
          "type": "image",
          "value": "Baker in 1940",
          "images": [
            {
              "content_url": "https://upload.wikimedia.org/wikipedia/commons/thumb/0/0b/Baker_Harcourt_1940_2.jpg/250px-Baker_Harcourt_1940_2.jpg",
              "caption": "Baker in 1940",
              "height": 250,
              "width": 250
            }
          ]
        },
        {
          "name": "Born",
          "type": "field",
          "value": "Freda Josephine McDonald June 3, 1906 St. Louis, Missouri, U.S.",
          "links": [
            {
              "url": "https://en.wikipedia.org/wiki/St._Louis",
              "text": "St. Louis"
            }
          ]
        },
        [...]

Article Sections

Parsed sections of the article body, as recursive trees of parts. The first section is named Abstract when it has no header

id=\"mwIQ\">During her early career, Baker was among the most celebrated performers to headline the revues of the <span title=\"French-language text\" about=\"#mwt45\" typeof=\"mw:Transclusion\" data-mw='{\"parts\":[{\"template\":{\"target\":{\"wt\":\"lang\",\"href\":\"./Template:Lang\"},\"params\":{\"1\":{\"wt\":\"fr\"},\"2\":{\"wt\":\"[[Folies Bergère]]\"},\"italic\":{\"wt\":\"no\"}},\"i\":0}}]}' id=\"mwIg\"><span lang=\"fr\" style=\"font-style: normal;\"><a rel=\"mw:WikiLink\" href=\"./Folies_Bergère\" title=\"Folies Bergère\">Folies Bergère</a></span></span><link rel=\"mw:PageProp/Category\" href=\"./Category:Articles_containing_French-language_text\" about=\"#mwt45\" id=\"mwIw\"/> in <a rel=\"mw:WikiLink\" href=\"./Paris\" title=\"Paris\" id=\"mwJA\">Paris</a>.[...]

\n\nDuring her early career, Baker was among the most celebrated performers to headline the revues of the {{lang|fr|[[Folies Bergère]]|italic=no}} in [[Paris]].[...]

"sections": [{
  "type": "paragraph",
  "value": "During her early career, Baker was among the most celebrated performers to headline the revues of the Folies Bergère in Paris. [...]",
  "links": [
    {
      "url": "https://en.wikipedia.org/wiki/Folies_Bergère",
      "text": "Folies Bergère"
    },
    [...]
  ],
  "citations": [
    {
      "identifier": "cite_note-4",
      "text": "[4]"
    },
    [...]
  ]
}]

Article Main Image

Main image of the article. Distinct from the images arrays inside content parts, which hold images used within the article body

<tr><td colspan=\"2\" class=\"infobox-image\"><span class=\"mw-default-size\" typeof=\"mw:File/Frameless\"><a href=\"./File:Baker_Harcourt_1940_2.jpg\" class=\"mw-file-description\"><img alt=\"Baker in 1940\" resource=\"./File:Baker_Harcourt_1940_2.jpg\" src=\"//upload.wikimedia.org/wikipedia/commons/thumb/0/0b/Baker_Harcourt_1940_2.jpg/250px-Baker_Harcourt_1940_2.jpg\" decoding=\"async\" data-file-width=\"540\" data-file-height=\"756\" data-file-type=\"bitmap\" height=\"350\" width=\"250\" srcset=\"//upload.wikimedia.org/wikipedia/commons/thumb/0/0b/Baker_Harcourt_1940_2.jpg/500px-Baker_Harcourt_1940_2.jpg 1.5x, //upload.wikimedia.org/wikipedia/commons/thumb/0/0b/Baker_Harcourt_1940_2.jpg/500px-Baker_Harcourt_1940_2.jpg 2x\" class=\"mw-file-element\"/></a></span><div class=\"infobox-caption\">Baker in 1940</div></td></tr> [...]

| image = File:Baker Harcourt 1940 2.jpg\n| caption = Baker in 1940\n| [...]

"image": {
  "content_url": "https://upload.wikimedia.org/wikipedia/commons/0/0b/Baker_Harcourt_1940_2.jpg",
  "width": 540,
  "height": 756
}

Article Lists

Values of the part, used by list parts

<section data-mw-section-id=\"1\" id=\"mwGw\"><h2 id=\"History\">History</h2>\n<ul id=\"mwHA\"><li id=\"mwHQ\">1964–1984: Wheelchair Powerlifting</li>\n<li id=\"mwHg\">1984–2016: Paralympic Powerlifting / IPC Powerlifting</li>\n<li id=\"mwHw\">2017–present: Para Powerlifting</li></ul>\n\n</section>

==History==\n* 1964–1984: Wheelchair Powerlifting\n* 1984–2016: Paralympic Powerlifting / IPC Powerlifting\n* 2017–present: Para Powerlifting\n\n

{
  "name": "History",
  "type": "section",
  "has_parts": [
  {
    "type": "list",
    "has_parts": [
    {
      "type": "list_item",
      "value": "1964–1984: Wheelchair Powerlifting"
    },
    {
      "type": "list_item",
      "value": "1984–2016: Paralympic Powerlifting / IPC Powerlifting"
    },
    {
      "type": "list_item",
      "value": "2017–present: Para Powerlifting"
    }]
  }]
}

Article Tables

Tables defined in the article, referenced from section parts by identifier. A table may be defined but never referenced, and a section may reference a table that is not present

[...]<table class=\"wikitable sortable\" id=\"mwBrw\">\n<caption id=\"mwBr0\">Film credits for Josephine Baker</caption>\n<tbody id=\"mwBr4\"><tr id=\"mwBr8\">\n<th scope=\"col\" id=\"mwBsA\">Year</th>\n<th scope=\"col\" id=\"mwBsE\">Title</th>[...]</tr>\n<tr id=\"mwBsc\">\n<th scope=\"row\" id=\"mwBsg\">1927</th>\n<td id=\"mwBsk\"><span title=\"French-language text\" about=\"#mwt805\" typeof=\"mw:Transclusion\" data-mw='{\"parts\":[{\"template\":{\"target\":{\"wt\":\"lang\",\"href\":\"./Template:Lang\"},\"params\":{\"1\":{\"wt\":\"fr\"},\"2\":{\"wt\":\"La Sirène des Tropiques\"}},\"i\":0}}]}' id=\"mwBso\"><i lang=\"fr\">La Sirène des Tropiques</i></span><link rel=\"mw:PageProp/Category\" href=\"./Category:Articles_containing_French-language_text\" about=\"#mwt805\" id=\"mwBss\"/> (<i id=\"mwBsw\"><a rel=\"mw:WikiLink\" href=\"./Siren_of_the_Tropics\" title=\"Siren of the Tropics\" id=\"mwBs0\">Siren of the Tropics</a></i>)</td>\n<td id=\"mwBs4\">Papitou</td>\n<td id=\"mwBs8\"><a rel=\"mw:WikiLink\" href=\"./Silent_film\" title=\"Silent film\" id=\"mwBtA\">silent film</a></td>[...]

[...]\n\n== Film credits ==\n{| class=\"wikitable sortable\"\n|+Film credits for Josephine Baker\n|-\n! scope=\"col\"| Year\n! scope=\"col\"| Title\n! scope=\"col\"| Role\n! scope=\"col\" class=\"unsortable\"| Notes\n! scope=\"col\" class=\"unsortable\"| {{abbr|Ref.|Reference}}\n|-\n!scope=row| 1927\n| {{lang|fr|La Sirène des Tropiques}} (''[[Siren of the Tropics]]'')\n| Papitou\n| [[silent film]]\n|align=\"center\" |{{sfnp|Bergfelder|Harris|Street|2007|p=193}}{{sfnp|Francis|2021|p=68}}\n|-\n!scope=row| 1927\n| {{lang|de|Die Frauen von Folies Bergères}} (''[[The Woman from the Folies Bergères]]'')\n|\n| [[silent film]]\n|align=\"center\" |[...]

"tables": [
  {
    "identifier": "film_credits_table1",
    "headers": [
      [
        { "value": "Year" },
        { "value": "Title" },
        { "value": "Role" },
        { "value": "Notes" },
        { "value": "Ref." }
      ]
    ],
    "rows": [
      [
        { "value": "1927" },
        { "value": "La Sirène des Tropiques (Siren of the Tropics)" },
        { "value": "Papitou" },
        { "value": "silent film" },
        { "value": "" }
      ],
      [
        { "value": "1927" },
        { "value": "Die Frauen von Folies Bergères (The Woman from the Folies Bergères)" },
        { "value": "" },
        { "value": "silent film" },
        { "value": "" }
      ],[...]
    ],
    "confidence_score": 0.8
  }
],

How to Access Structured Contents

Structured Contents is currently available in two of our APIs:

On-demand API: request individual articles from any project with structured JSON. Best for testing, post-training, or lightweight use. Follow the On-demand guide, or jump to the endpoint reference.

Snapshot API: get a compressed file of all articles in a project as structured JSON snapshots. Best for pre-training, indexing, and high-scale applications. Follow the Snapshot guide, or jump to the endpoint reference.

Narrow each response to just the parts you need with Fields, Filters, and Limit. A free account covers both APIs; talk to Sales for more requests, or data updated more often than monthly.

Get Started for free

Contact Sales

Wikimedians can also access beta Structured Contents through their Wikimedia Cloud Services accounts.

Shaping Structured Contents Together

We continue to build on top of recent Structured Contents releases, in response to feedback you've given us: we're improving our recent updates to tables and references so they are parsed more in line with user expectations. We're also working to expand coverage to more Wikipedia languages.

In order to help us strengthen current features and shape new ones, we welcome and encourage feedback on Structured Contents. If you have a use case that warrants adding support for a new language to our Structured Contents coverage, please let us know.

The On-demand and Snapshot APIs are how you build on Structured Contents, and a free account already covers both. For the Hugging Face and Kaggle communities, we have also shared French and English Structured Contents snapshots in Parquet format, so you can experiment with the data in tools you already use.

Hugging Face Dataset →

Kaggle Dataset →

Structured Contents Payload Example

Our Structured Contents endpoints have the same familiar structure as our production responses, but also include beta fields and objects parsed from raw article data. Parsed objects that are unique to Structured Contents are: infoboxes, sections, description, references, and tables.

Structured Contents responses use the structured_content model, not the production article model. Alongside what it adds, know what it leaves out: the response does not carry article_body, categories, templates, redirects, protection, visibility, watchers_count, or the previous_version / date_previously_modified change-history fields. When you need those, call the production endpoints and join the two responses on identifier. Structured Contents endpoints are beta: experimental and not covered by SLA.

Example: Run this cURL command with your access token (see auth docs) to get the Structured Contents response from the live English Josephine Baker Wikipedia article as seen here →

curl --location 'https://api.enterprise.wikimedia.com/v2/structured-contents/Josephine_Baker' --header 'Content-Type: application/json' --header 'Authorization: Bearer ACCESS_TOKEN' --data '{"filters":[{"field":"is_part_of.identifier","value":"enwiki"}]}'

For a step-by-step version of this lookup, see the worked examples in the On-demand API guide. For a full breakdown and explanation of all Structured Contents response fields, consult the structured_content model in our API reference.

Keep exploring

Recent releases and deep-dives from the Structured Contents Initiative:

See all Structured Contents articles →