Search

Search runs entirely in the browser. There is no server and no external service.

How it works

When tago builds the site, it writes a search index file, en.search-data.json, with one entry per page: the page's title, its link, and its text capped to a few hundred characters. The theme loads FlexSearch and indexes that file in the browser. As the reader types in the navbar box, matches appear in a dropdown.

There is also a full-page search view. Create a page with the search layout and it renders the same results in the page body:

---
title: "Search"
layout: "search"
ExcludeSearch: true
---

What is indexed

Every page is indexed except those that set ExcludeSearch: true in front matter. Set it on the search page itself so it does not list itself as a result. The indexed text is the rendered page content, so headings and prose are searchable; front matter and code are weighted by their place in the body.

Search and sub-path deploys

The index links each result by its site-root path, without any deploy sub-path. When the site is served under a sub-path, for example username.github.io/project/, the theme reads the configured base URL and prepends its path to each result link, so search results resolve correctly in both a root deploy and a sub-path deploy. You do not configure this; it follows the base URL you build with. See building for production for how the base URL is set.