Skip to main content
Connect a WordPress site and Raily reads its content through the standard WordPress REST API, indexes it, and keeps it in sync. You don’t install a plugin on your site.

Before you start

  • Your site’s REST API must be reachable from the public internet at https://yoursite.com/wp-json. It is on by default in WordPress. Some security plugins and firewalls turn it off or block it.
  • You need permission to manage resources in your Raily Organization. Without it, the Connect source button doesn’t appear.
  • Optional: a WordPress username and an Application Password, if your site only answers the REST API for signed-in users.

Connect the site

1

Open the source catalog

In the Rail, click Connect source. The Connect a data source page opens.
2

Pick WordPress

Click WordPress (listed under CMS).
3

Name the source

In Name, type a name you’ll recognize later, such as Company blog.
4

Enter the site address

In Site URL, enter the public address of your site, for example https://yourblog.com. If WordPress is installed in a subfolder, include it (https://example.com/blog).
5

Add credentials (only if needed)

Public content needs no login. If your site requires one, click Add access under Private content, then fill in WordPress username and Application Password. You create an Application Password in WordPress under Users → Profile → Application Passwords. It’s not your login password, and you can revoke it at any time.
6

Test the connection

Click Test connection. Raily checks the site, verifies the credentials if you added them, and lists every type of content the site publishes.
7

Choose what to make searchable

Under What should we make searchable?, keep or clear the content types. After a successful test, every type the site publishes is selected: Posts, Pages, and any custom types your site uses (many sites keep most of their content in a custom type, such as articles or reports). You can select up to 24 types.
8

Set a language (optional)

In Language (optional), enter a language code such as en, or leave it blank.
9

Connect

Click Connect WordPress. The button stays disabled until a connection test succeeds.

What happens next

Raily opens the new Source and starts profiling your content. Once indexing finishes, Raily generates the search configuration for the site’s collection, and you can wire it to an endpoint. Each WordPress Source holds one collection for the whole site. The Source’s sync card shows progress while the first sync runs, then the number of items added, updated, and removed. After that, Raily checks the site for new, changed, and removed content about once an hour. To check right away, click Resync on the sync card. If content looks wrong, open the menu next to Resync and choose Full reindex to re-read and re-process every item.

What gets indexed

For each item in the content types you selected, Raily indexes:
  • The title and body text
  • The author (the item’s WordPress author)
  • The publish date and last-modified date
  • Categories, tags, and the excerpt
  • The content type (post, page, or a custom type)
  • The featured image, or the first image in the body if there is no featured image
  • A link to the item on your site
Only published items are indexed. Search results link back to the item’s page on your site. Some page builders render the page from a template, so WordPress returns an empty body for it. When that happens, Raily reads the main content area of the live page instead.

Linked PDFs

Many sites publish articles whose real content is a PDF: an archive article that is just an embedded PDF viewer, or a short summary with a Download the PDF link. Raily reads those documents and indexes their text as part of the article. A document is included when the article body contains it as:
  • An embedded viewer (a PDF shown inside the page), including viewers that wrap the PDF address
  • A link to the file, including a WordPress file block
The document must:
  • Be a PDF or a Word document (.docx)
  • Live on your own site. Links to PDFs on other websites, such as a report you cite, are treated as citations and are not indexed. Your site with and without www. counts as the same site.
Raily reads up to 5 documents per article. When an article has its own text as well, Raily indexes the article text first, followed by the text of each document in the order they appear. The result is still one searchable item: the title, author, date, categories, and link all come from the WordPress article, not from the file. A search that matches text inside the PDF returns the article. Scanned PDFs are supported. Raily extracts their text, which takes longer than for a PDF with a text layer.

Limits

  • If a document can’t be downloaded (for example, it was deleted) or can’t be read, Raily skips it and indexes the article with its own text.
  • If you replace a PDF with a new file at the same address without editing the article, Raily keeps the old text. Edit the article, or run Full reindex, to pick up the new file.

Edit or remove the source

Open the Source and edit it to change the content types, language, or credentials, or to turn it on or off. When you change credentials, click Test connection again before saving. The Site URL can’t be changed after you connect, because it identifies the Source. To index a different site, connect a new WordPress Source. To remove the Source, click Delete source.

Next steps

Create an AI endpoint

Let agents search this source

Data Catalog

Review the fields agents filter on