Documentation

Plain fetch

No SDK, no build step. This is the whole integration for a list of posts.

const response = await fetch(
  "https://api.schemahaus.com/v1/content/blog-post?locale=ka&limit=10",
  { headers: { authorization: `Bearer ${process.env.SCHEMAHAUS_TOKEN}` } },
);

const { data, meta } = await response.json();
// data is only what is PUBLISHED in Georgian. Nothing is substituted
// unless you ask: add &fallback=chain, and check item.localeFallback.

Next.js

The SDK plus a revalidation webhook, so an editor pressing publish rebuilds the page rather than waiting for a cache to expire.

// app/[locale]/blog/[slug]/page.tsx
import { createClient } from "@schemahaus/client";

const cms = createClient({
  token: process.env.SCHEMAHAUS_TOKEN,
  site: process.env.SCHEMAHAUS_SITE,
});

export default async function Post({ params }) {
  const { locale, slug } = await params;
  const post = await cms.document("blog-post", slug, { locale });

  if (!post) notFound();
  return <article dangerouslySetInnerHTML={{ __html: post.data.body }} />;
}

// Publishing fires a webhook. Point a nextjs_revalidate webhook at your
// deploy and the page rebuilds itself when an editor presses publish.

Jekyll, at build time

For a static site that has no runtime at all: pull the content during the build and write the files the generator expects.

# _plugins/schemahaus.rb — runs at build time, writes _posts/
require "net/http"
require "json"

site = ENV.fetch("SCHEMAHAUS_SITE")
token = ENV.fetch("SCHEMAHAUS_TOKEN")

uri = URI("https://api.schemahaus.com/v1/content/blog-post?locale=en")
request = Net::HTTP::Get.new(uri)
request["authorization"] = "Bearer #{token}"

response = Net::HTTP.start(uri.hostname, uri.port, use_ssl: true) do |http|
  http.request(request)
end

JSON.parse(response.body)["data"].each do |post|
  File.write("_posts/#{post["publishedAt"][0, 10]}-#{post["slug"]}.md",
             "---\ntitle: #{post["data"]["title"].inspect}\n---\n" +
             post["data"]["body"])
end

Delivery

Published content, read with a delivery token. Safe to put in a consumer site's environment: it can read and nothing else.

• GET/api/v1/authors

  The site's authors, with their per-language biographies.

• GET/api/v1/content/{apiId}

  Published documents of one content type. Filter, sort, paginate by `page` or by cursor (`?cursor=` to start, then each page's `meta.nextCursor`); `?locale=` is strict unless `fallback` says otherwise.

• GET/api/v1/content/{apiId}/{identifier}

  One published document, by slug or by document id.

• GET/api/v1/media/{id}

  One media asset's metadata and variant URLs.

• POST/api/v1/preview/verify

  Exchange a preview token for the draft it was minted for. Needs a delivery token with draft access.

• GET/api/v1/site

  The site's settings and its content languages.

• GET/api/v1/site/types

  The content types on the site, with their fields. Addressed by apiId; no internal ids are returned.

Management

Writes, bound to a membership, so a key can never do more than the person who minted it. This is what the import scripts use.

• GETPOST/api/v1/manage/authors

  List or create authors.

• PATCHDELETE/api/v1/manage/authors/{id}

  Edit or remove one author.

• POST/api/v1/manage/documents

  Create a document in one language.

• PATCHDELETE/api/v1/manage/documents/{id}

  Edit one language version, or soft-delete the whole document. Editing takes the revision you last read.

• POST/api/v1/manage/documents/{id}/locales

  Start a translation of a document.

• POST/api/v1/manage/documents/{id}/transition

  Move one language version through the workflow, including publishing it.

• GET/api/v1/manage/documents/lookup

  Find a document by slug, drafts included — how an import decides whether it has already brought a post across.

• POST/api/v1/manage/media

  Reserve an upload and get a presigned PUT for object storage.

• POST/api/v1/manage/media/{id}/complete

  Confirm the bytes arrived. Idempotent, so a retried import does not count the storage twice.

• GET/api/v1/manage/types

  Content types WITH their internal ids, which every write below needs.

Strapi compatibility

Strapi's query shape over the same published content, so an existing site can move before it is rewritten.

• GET/api/strapi/api/{plural}

  The Strapi compatibility layer: Strapi's query shape over the same published content, so an existing site can move before it is rewritten.