ARTICLES.SANTISMM.COM

Articles API

Read the essays as data.

Everything on this site is also published as structured, read-only data. No key, no account, no SDK: plain HTTPS GET requests for static files.

Endpoints

All URLs are relative to https://articles.santismm.com. The machine-readable description is /openapi.json (OpenAPI 3.1); every operation in it has an operationId.

PathFormatWhat it is
/api/articles.jsonJSONEvery essay with metadata and full text, newest first.
/api/articles/{slug}.jsonJSONOne essay; identical to its entry in the catalogue.
/ai-index.jsonJSONMetadata for every essay (no text) and the URL of every discovery surface.
/feed.json, /feed.{en,es,pt}.jsonJSON Feed 1.1All essays, or one language.
/feed.xml, /feed.{en,es,pt}.xmlRSS 2.0All essays, or one language.
/llms.txttext/plain (Markdown)Agent index with guidance on when to use and cite this corpus.
/llms-full.txttext/plain (Markdown)The full text of every essay in one document.
/HTML or MarkdownThe index. Send Accept: text/markdown for the Markdown rendition.
/openapi.jsonOpenAPI 3.1Machine-readable description of the JSON endpoints.
/sitemap.xmlXMLEvery page, with language alternates.

For example, /api/articles/a-empresa-agentica-precisa-de-um-sistema-imunologico.json is the essay A empresa agêntica precisa de um sistema imunológico.

Formats

  • JSON is UTF-8, pretty-printed, served as application/json; charset=utf-8.
  • Every Articles API document carries "schema_version": "1.0" and a license object.
  • An essay's body is Markdown-like plain text derived from the authored HTML: # headings, - list items, links written as text (url).
  • Dates are YYYY-MM-DD. language is a BCP 47 tag: English (en), Español (es-ES), Português (pt-BR).
  • Language editions of the same essay share a translation_key; it is absent on essays with no translation.
  • Cite canonical_url, never the API URL.

Content negotiation

  • GET / with Accept: text/markdown returns a Markdown rendition of the index (text/markdown; charset=utf-8); any other Accept returns HTML. Both responses carry Vary: Accept.
  • Essay pages are HTML only. Their full text is available as JSON at /api/articles/{slug}.json and in /llms-full.txt.

Errors

The only error a client should meet is 404 for a path that does not exist. Its body follows the Accept header and the response carries Vary: Accept:

  • application/json, or any unknown path under /api/ unless the client prefers HTML: {"error":{"code":"404","message":"The page could not be found","path":"…","documentation_url":"…"}}
  • text/markdown: a short Markdown page pointing back to the index and this documentation.
  • Anything else: the hosting platform's own plain-text or HTML 404 page.

Methods other than GET, HEAD and OPTIONS are answered 405 by the host. Query strings are ignored.

Versioning

There is one version, schema_version "1.0", and no version in the paths. The contract is additive: new essays, new fields and new endpoints can appear at any time, so clients must ignore fields they do not know. A field is never removed, renamed or retyped in place — a breaking change would be published under a new path, with schema_version raised, and the current paths would keep their current shape.

Caching and CORS

  • Responses under /api/ send Cache-Control: public, max-age=0, s-maxage=300, stale-while-revalidate=86400 and an ETag; conditional requests with If-None-Match get 304 Not Modified.
  • /api/ and /openapi.json send Access-Control-Allow-Origin: *, so they can be read from any origin in a browser.

Rate limits

None is enforced by this site: there is no key, quota or rate-limit header. The files are served by Vercel, whose platform-level abuse protection applies and is not configured or documented by this site. Be considerate: /api/articles.json holds the whole corpus in one request, and the feeds tell you when anything changed.

Authentication and licence

None required. The essays are licensed CC BY 4.0: attribute Santiago Santa María Morales and cite each essay's canonical URL. santismm.com consumes this API for the article tools of its MCP endpoint; this site remains the source of truth.

What is not here

No search, filtering, pagination or write operation; no MCP endpoint on this host; no per-essay Markdown endpoint other than the JSON body.