View source for EveryCRT:Data API

{{TOC right}}
The '''EveryCRT Data API''' is a documented way to fetch the structured data of [[Main Page|EveryCRT]] programmatically — for example, to build a third-party app that mirrors, indexes, or analyzes the wiki.

It is not a custom REST service. It is a thin documentation layer over the standard '''MediaWiki Action API''' and '''[https://www.semantic-mediawiki.org/ Semantic MediaWiki]''' endpoints, plus one wiki-maintained [[#Manifest|manifest]] that lists the canonical entity-type categories.

No authentication is required. All data is publicly readable.

== Sync pattern ==

The recommended way to mirror the wiki is two-step:

# Fetch the '''[[#Manifest|manifest]]''' once to discover the entity types and how to enumerate each.
# For each entity type, fetch its '''[[#Listing entities|list of pages]]''' (with last-modified timestamps), then call '''[[#Entity facts|entity facts]]''' on each page you want full data for.

For ongoing incremental sync, poll '''[[#Recent changes|recent changes]]''' since your last sync timestamp and re-fetch only the affected pages.

== Manifest ==

The manifest is a JSON document listing every canonical entity-type category, with live page counts and ready-to-use enumeration URLs. It updates automatically as the catalog grows; new entity types added in the future will appear here without API clients needing code changes (assuming they iterate over <code>entity_types</code> rather than hardcoding).

'''Human-readable view:''' [[EveryCRT:Data API/manifest]]

'''Programmatic fetch (recommended):'''

<pre>GET https://wiki.everycrt.com/w/api.php?action=expandtemplates&text=%7B%7B%23invoke%3ADataAPIManifest%7Cjson%7D%7D&prop=wikitext&format=json&formatversion=2</pre>

The response wraps the manifest in <code>expandtemplates.wikitext</code> as a string; parse that string as JSON:

<syntaxhighlight lang="javascript">
const url = 'https://wiki.everycrt.com/w/api.php'
  + '?action=expandtemplates'
  + '&text=%7B%7B%23invoke%3ADataAPIManifest%7Cjson%7D%7D'
  + '&prop=wikitext&format=json&formatversion=2';

const res = await fetch(url, {
  headers: { 'User-Agent': 'my-app/1.0 (https://example.com/contact)' },
});
const { expandtemplates } = await res.json();
const manifest = JSON.parse(expandtemplates.wikitext);

console.log(manifest.entity_types);
</syntaxhighlight>

== Listing entities ==

Use Semantic MediaWiki's <code>action=ask</code> endpoint to enumerate pages of a given entity type. The manifest's <code>list_url</code> for each entity type does exactly this — returning every page with its <code>Modification date</code>, which is what you want for indexing and for driving incremental sync:

<pre>GET https://wiki.everycrt.com/w/api.php?action=ask&format=json&formatversion=2&query=[[Category:CRT models]]|?Modification date|limit=500</pre>

The query supports any SMW ask syntax — additional printouts, filters, ordering, etc.

* Page through with <code>|offset=N</code> in the query.
* Maximum <code>limit</code> per request is 500.

'''Example — all Sony CRT models with screen size and release year:'''

<pre>GET https://wiki.everycrt.com/w/api.php?action=ask&format=json&query=[[Category:CRT models]][[Has brand::Sony]]|?Has screen size inches|?Has release year|limit=500</pre>

For interactive query development, use [[Special:Ask]].

== Entity facts ==

For a single page, the <code>action=browsebysubject</code> endpoint returns ''every'' SMW fact known about it — including subobjects (CRT inputs, outputs, voltage ranges, etc.):

<pre>GET https://wiki.everycrt.com/w/api.php?action=browsebysubject&format=json&formatversion=2&subject=Sony_KV-20FV300</pre>

This is the right endpoint for "give me everything you know about this entity" — no need to enumerate properties up front. New properties added to the wiki are automatically included in the response.

Subobjects are returned as additional subjects with names like <code>Sony_KV-20FV300#_HASH</code>. Treat them as child records of the parent page; together they describe the CRT inputs, outputs, voltage ranges, and so on.

For interactive browsing, use [[Special:Browse]].

== Recent changes ==

For incremental sync, poll <code>action=query&list=recentchanges</code> filtered by a timestamp. The standard MediaWiki [[Special:RecentChanges]] view is the human-readable equivalent:

<pre>GET https://wiki.everycrt.com/w/api.php?action=query&list=recentchanges&format=json&formatversion=2&rcprop=title|timestamp|ids|sizes|user&rclimit=500&rcend=2026-05-01T00:00:00Z</pre>

A typical sync loop:

# Save your "last sync" timestamp (UTC ISO 8601) after each successful pass.
# On the next sync, fetch recent-changes with <code>rcend=<previous timestamp></code>.
# For each unique <code>title</code> in the response, re-fetch via <code>browsebysubject</code> (or remove from your index if it was a deletion — see <code>rctype=log</code> and <code>rcshow</code> options for handling deletes and moves).
# Save the new "last sync" timestamp.

To filter to specific namespaces, add <code>&rcnamespace=0|3000</code> (main + Document, for example). Namespace IDs are listed via:

<pre>GET https://wiki.everycrt.com/w/api.php?action=query&meta=siteinfo&siprop=namespaces&format=json&formatversion=2</pre>

== RDF export ==

For semantic-web-style consumption, [[Special:ExportRDF]] returns RDF/OWL for individual pages or entire category subtrees:

<pre>GET https://wiki.everycrt.com/wiki/Special:ExportRDF?pages=Category:CRT%20models&recursive=1</pre>

This is more verbose than the JSON endpoints; most app developers will prefer <code>browsebysubject</code>.

== Conventions and stability ==

* All endpoints return UTF-8.
* Page titles are case-sensitive after the first character (standard MediaWiki).
* '''Subobjects''' (per-input, per-output, per-voltage-range facts) appear as separate subjects in <code>browsebysubject</code> output, with titles ending in <code>#_HASH</code>. Treat them as child records of the parent page; you do not need to fetch them separately.
* The set of properties on any one entity type may grow over time. Clients using <code>browsebysubject</code> automatically pick up new properties; clients using fixed <code>action=ask</code> printouts will need updates when they want to read newly added fields.
* Entity-type categories listed in the manifest are stable. New entity types may be added; existing ones will not be silently renamed.
* This is a community wiki; data quality varies. Treat values as best-effort, especially for newer or less-documented models.

== Etiquette ==

The wiki is publicly readable and no authentication is required for any of these endpoints. Please be a good citizen:

* Set a descriptive <code>User-Agent</code> on your requests (e.g. <code>my-app/1.0 (https://example.com/contact)</code>) so the wiki operators can reach you if there's a problem.
* Cache the manifest — it changes slowly (new entity types are rare).
* For full syncs, sleep briefly between requests; a few requests per second is fine.
* Use [[#Recent changes|recent changes]] for incremental sync rather than re-walking the whole catalog every time.

If your app is doing something cool with EveryCRT data, drop a note on the community channels — we'd love to see it.

== See also ==

* [[Special:RecentChanges]] — human-readable change log
* [[Special:Ask]] — interactive SMW query builder
* [[Special:Browse]] — interactive per-page fact browser
* [[EveryCRT:Data API/manifest]] — human-readable view of the manifest
* [[Module:DataAPIManifest]] — Lua source of the manifest
* [[EveryCRT:Maintenance]] — wiki health and category audits

[[Category:EveryCRT administration]]