WordPress REST API Scraper: Export Posts & Pages to Markdown

WordPress REST API scraper and export tool: get the posts and pages of a WordPress site as JSON, HTML and Markdown, with categories, tags, authors and images. No login needed.

WordPress REST API Scraper: Export Posts & Pages to Markdown

Give it one or more WordPress sites and get every post and page as clean rows: title, slug, URL, publish and modified dates, author name, category and tag names, plain-text excerpt, full content as HTML and as Markdown, featured image URL and word count. It reads the same public JSON endpoints that WordPress itself publishes under /wp-json/wp/v2/, so no API key, no browser and no proxy are needed.

It needs the public REST API. Sites that switched it off, protect it with a login or firewall, or forbid /wp-json/ in robots.txt cannot be exported; for those you get one free status row that says why.

• Two content formats: contentHtml for exact reproduction, contentMarkdown ready for static-site generators, notes tools and language-model pipelines.
• Filters: content types (posts, pages), date range, categories and tags (slugs or IDs), and a maximum number of items per site.
• Several sites per run: each site is read at about one request per second and its robots.txt is respected.
• Never crashes on a closed site: when the REST API is switched off, blocked or missing, you get one free status row that says why (HTTP 401/403/404, an HTML page instead of JSON, or a robots.txt rule).
• Email addresses are removed from author names, excerpts and exported article text. Other personal data in published content may remain.

Use cases

• WordPress export and site migration: turn WordPress posts into Markdown (WP posts to Markdown) and move a blog or a documentation site to another CMS or a static site generator, with title, slug, dates, categories and tags as front-matter data.
• Content audit and SEO: list every URL with title, excerpt, word count, categories, tags and last-modified date to find thin, outdated or untagged content.
• LLM datasets: build clean text corpora, retrieval (RAG) knowledge bases and fine-tuning sets from sites you are allowed to use, with Markdown that keeps headings, lists and links.
• Backup and archive: keep a dated copy of everything a WordPress site published, as JSON, CSV or Excel from the dataset.

Input

Field	Description
sites	WordPress domains or URLs, for example wptavern.com or https://example.com/blog (sub-directory installs work).
postTypes	posts, pages or both. Default both; posts come first.
maxItemsPerSite	Posts and pages together, posts first, newest first within each type. Default 100: a site with 100 or more posts returns no pages unless you raise this or choose only "pages". 0 = the whole site.
dateFrom, dateTo	Optional publish date range (YYYY-MM-DD, site time zone, both ends included).
categories, tags	Optional slugs or numeric IDs. Only posts are filtered; when set, pages are left out.
includeContent	Default on. Turn off for a light list without contentHtml, contentMarkdown and wordCount.
includeSiteInfo	Default on. One free site row per site.
maxConcurrency	Sites read in parallel (default 2). Each single site still gets about one request per second.

{
  "sites": ["wordpress.org/news", "wptavern.com", "https://developer.wordpress.org/news/"],
  "maxItemsPerSite": 20
}

Output

One row per post or page. Long text is shortened with "…" in this example.

{
  "type": "post",
  "site": "developer.wordpress.org/news",
  "id": 6415,
  "title": "What’s new for developers? (October 2026)",
  "slug": "whats-new-for-developers-october-2026",
  "url": "https://developer.wordpress.org/news/2026/10/whats-new-for-developers-october-2026/",
  "date": "2026-10-14T19:56:09Z",
  "modified": "2026-10-14T19:56:25Z",
  "author": "Justin Tadlock",
  "categories": ["Blocks", "Playground", "Plugins", "Themes"],
  "tags": ["Monthly Roundups"],
  "excerpt": "Your October guide to WordPress development: the 7.2 roadmap, the Ipsum theme, new block APIs, and what to test before Beta 1.",
  "contentHtml": "\n<p class=\"wp-block-paragraph\">This edition is running a bit late. I was sick last week, …",
  "contentMarkdown": "This edition is running a bit late. I was sick last week, and it seems that sorting through two Gutenberg releases and a…",
  "wordCount": 2328,
  "featuredImageUrl": "https://developer.wordpress.org/news/files/2026/10/whats-new-devs-oct-2026.png",
  "input": "https://developer.wordpress.org/news/",
  "scrapedAt": "2026-10-23T20:41:35.482Z"
}

type is post or page. Pages have no categories and tags. Dates are UTC. wordCount counts words of the plain text of the body.

Free helper rows share the same dataset:

• type: "site": site, name, description, homeUrl, timezone, restApi, postsMatching, pagesMatching, itemsSaved, complete (false when an item/budget limit or a later-page failure stopped the export; partial rows remain counted in itemsSaved), authorNamesAvailable.
• type: "status": status, input, httpStatus, message.

Status values you may see: rest_api_disabled (HTTP 401/403), rest_api_not_found (HTTP 404 or an HTML page instead of JSON), blocked_by_robots, fetch_failed, category_not_found, tag_not_found, empty, invalid_url, invalid_input.

{
  "type": "status",
  "status": "rest_api_disabled",
  "input": "css-tricks.com",
  "message": "The REST API answered HTTP 401: it is disabled or blocked for public visitors. Not charged.",
  "site": "css-tricks.com",
  "httpStatus": 401,
  "scrapedAt": "2026-10-23T20:44:03.042Z"
}

Limits

• Only public HTTP(S) sites without URL credentials are supported. Local/private/reserved IP addresses and mixed public/private DNS answers are rejected; each connection is pinned to a vetted public address. Redirects (up to five) and advertised REST roots follow the same checks, and each content hop must pass its target site's robots rules. Responses larger than 8 MiB (512 KiB for robots.txt) are stopped and reported as free failure rows. Earlier saved post/page rows remain counted results and partial site summaries say complete: false; exporting every site regardless of response size is not guaranteed.
• The site's public REST API must be on. Sites that disable or protect it, or that forbid /wp-json/ in robots.txt, return a status row only.
• Only published, public posts and pages are exported. No drafts, no private or password-protected bodies, no custom post types, no comments.
• If a site hides its users endpoint, author is empty (the site row shows authorNamesAvailable: false).
• Date filters use the site's own time zone, as WordPress does.
• Email addresses are removed from exported excerpts and article text. Other personal data in published content may remain.

FAQ

Does it work with every WordPress site? It works with every site that leaves the standard REST API (/wp-json/wp/v2/) open, which is the default for WordPress. Sites using security plugins, firewalls or a headless setup may close it. In that case the run still succeeds and you get one free status row per site that tells you the reason.

How fast is it, and is it polite to the site? About one request per second per site, with retries and back-off on errors, and the site's robots.txt is read first and obeyed. For example, a site that disallows /wp-json/ for all crawlers is skipped with a blocked_by_robots row instead of being fetched.

Does it collect personal data? Author e-mail addresses are removed, as are e-mail addresses in excerpts and article text. Other personal data in published content may remain; please check the site's terms and the laws that apply to you before you reuse its articles.

How do I export only part of a site? Set dateFrom and dateTo, list category or tag slugs, choose posts or pages, and cap the size with maxItemsPerSite. The site row tells you how many items matched and whether your limit cut the export short.

Why is featuredImageUrl empty for some posts? The post has no featured image, or the site hides its media endpoint.

Pricing

Pay per result — see the Pricing tab.

Each saved post or page row is one result. Status rows and site info rows are free. The run stops when it reaches the maximum cost per run that you set, including across concurrently processed sites. With no remaining result budget it makes no requests. On graceful abort it saves RUN_SUMMARY and a STATE snapshot of saved site/type/item IDs before shutdown; snapshots do not automatically resume a later run.

Support and disclaimers

Questions and feature requests belong on the Issues tab of this Actor's page on Apify. This Actor is an independent tool and is not affiliated with WordPress, Automattic or the sites it reads. It reads only what a site publishes through its public REST API; how you may use that content depends on the site's terms and the applicable law.