Fixing Trailing-Slash Inconsistency

Why the slash creates two pages

A URL path is just a string. Nothing in the HTTP rules says that a path ending in a slash and the same path without one must lead to the same thing. Historically the slash meant "this is a folder" and no slash meant "this is a file", which is why a traditional server asked for /team, where team is a folder, answers with a redirect to /team/ and then serves the index.html inside it. Modern sites rarely map URLs to real folders, so the router in the CMS or framework decides what each form does, and different routers decide differently.

There are three states a URL can be in:

Google treats the two forms as separate URLs, and it will usually pick one as canonical by itself using the signals it has: redirects, canonical tags, internal links and the sitemap. When those signals agree, a stray duplicate causes very little trouble. When they disagree, for example the menu links without a slash, the sitemap lists slashes, and both forms return 200, you leave the choice to Google and may see the "wrong" form in results or reports split across two addresses.

Finding out what your site does now

Start with evidence rather than assumptions, because different parts of one site often behave differently. Blog posts may be handled by the CMS, product pages by a shop plugin and landing pages by static files, each with its own slash habit.

Pick ten to twenty URLs covering each kind of page: the homepage, a top-level page, a nested page, a blog post, a category or tag page, a product, a search results page, a PDF and a real folder if you have one. For each, request both forms without following redirects:

for u in /pricing /pricing/ /blog/spring-menu /blog/spring-menu/ /files/menu.pdf /files/menu.pdf/; do
  printf "%-28s " "$u"
  curl -s -o /dev/null -w "%{http_code} %{redirect_url}\n" "https://www.example.co.uk$u"
done

Typical output from a site that has drifted looks like this:

/pricing                     200
/pricing/                    200
/blog/spring-menu            301 https://www.example.co.uk/blog/spring-menu/
/blog/spring-menu/           200
/files/menu.pdf              200
/files/menu.pdf/             200

That tells you three things: the blog already has a with-slash convention, ordinary pages have none, and something is happily serving a PDF at an address ending in a slash, which usually means a catch-all route is swallowing it. Now run the same URLs through the Redirect Checker from the http and non-www starting points as well, because the slash rule usually interacts with the host rules. Finally, look at Search Console's page indexing report and the URL Inspection tool for a couple of pages to see which form Google has chosen as canonical. That is useful evidence when you choose.

Choosing with or without

Neither form ranks better. The deciding factors are practical: which form your existing links use, which form your platform produces naturally, and how much work each choice creates.

• Follow the majority of existing links. Check which form appears in your sitemap, your menus and, if you can see them, your backlinks. Switching thousands of established URLs to the other form creates redirects you did not need.
• Follow the platform. WordPress with the common "post name" permalink setting produces a trailing slash, because the structure ends in one. Many static site generators write each page as folder/index.html, which a plain web server serves at the slash form. Next.js defaults to no trailing slash unless its trailingSlash option is switched on. Fighting a platform's default usually means maintaining extra code forever.
• Consider files. Anything with an extension, such as .html, .pdf, .xml, .jpg or .css, should never gain a slash. If your site uses .html page addresses, the question is already answered for those pages.
• Consider forms and APIs. A 301 or 302 on a POST request lets browsers change the request into a GET, which drops the submitted data. If a form or API endpoint sits behind a slash rule, either exclude it or use a 308, which keeps the method.

Once you have chosen, write the decision down in one sentence, for example "All page URLs end in a slash except files with an extension", and make every later change follow it.

Where the slash behaviour comes from

WordPress

The permalink structure under Settings, Permalinks controls it. If the structure ends in /, WordPress generates slash links and its built-in canonical redirect sends the no-slash form to the slash form; remove the final slash and it does the reverse. Problems appear when a redirect plugin or an .htaccess rule enforces the opposite of the permalink setting, which produces a loop, or when pages created by a shop or membership plugin use their own routes. Change the permalink structure only as a deliberate project, because it changes every URL on the site.

Apache

The mod_dir module's DirectorySlash setting, on by default, redirects a request for a real folder without its slash to the slash form. That is correct for real folders and is why switching it off is rarely wise: with it off, relative links inside folder index pages break and directory listings can leak. Rules in .htaccess can add or remove slashes for everything else.

nginx

A try_files $uri $uri/ ... line makes nginx treat a request that matches a folder as a directory request, and nginx then issues its own 301 to add the slash. Behind a load balancer or proxy that redirect can point to the wrong scheme or port, such as http://www.example.co.uk:8080/team/, which is a classic source of odd extra hops. Setting absolute_redirect off; makes nginx send a relative Location instead, and port_in_redirect off; stops it adding the internal port.

Frameworks and static hosts

Next.js, Nuxt, Astro, Gatsby, Hugo, Eleventy and similar tools each have an option or a convention for this; Next.js, for example, redirects the non-preferred form with a 308 by default. Static hosts such as GitHub Pages, Netlify and Cloudflare Pages also normalise URLs according to how the files are laid out, and some offer a "pretty URLs" style setting. Check the documentation for the exact version you run and let one of these layers own the rule, not two.

Shopify and other hosted builders

On fully hosted platforms you usually cannot change the slash behaviour at all. Focus on the parts you do control: use the form the platform produces in every link you write, in campaigns and in any external listings.

Rules that handle files and folders correctly

The rule has to do three things: act only on paths that are not real files, leave real folders to the server's own behaviour, and send the visitor straight to the final HTTPS host in the same hop. Here are tested patterns for the two common choices.

Apache, removing the slash

RewriteEngine On
# leave real folders alone; mod_dir handles them
RewriteCond %{REQUEST_FILENAME} !-d
# only paths that end in a slash and are not the root
RewriteCond %{REQUEST_URI} ^(.+)/$
RewriteRule ^ https://www.example.co.uk%1 [R=301,L]

Apache, adding the slash

RewriteEngine On
RewriteCond %{REQUEST_FILENAME} !-f
# skip anything that looks like a file name with an extension
RewriteCond %{REQUEST_URI} !\.[a-zA-Z0-9]{1,5}$
RewriteCond %{REQUEST_URI} !/$
RewriteRule ^ https://www.example.co.uk%{REQUEST_URI}/ [R=301,L]

nginx, removing the slash

# inside the server block for the canonical host
location ~ ^(.+)/$ {
    if (!-d $request_filename) {
        return 301 $scheme://$host$1$is_args$args;
    }
    try_files $uri $uri/ /index.php?$args;
}

nginx, adding the slash

rewrite ^([^.?]*[^/])$ $1/ permanent;

The nginx add-slash rewrite only matches paths with no dot and no final slash, so /menu.pdf and /sitemap.xml are left alone. rewrite keeps the query string automatically. Place slash rules after your host and HTTPS handling or merge them into it, as the next section explains, and test with nginx -t or apachectl configtest before reloading.

The fix, step by step

Edge cases worth checking

• Query strings. /search?q=shoes and /search/?q=shoes are separate too. Make sure your rule keeps the query string when it redirects, or campaign tags and search terms are lost.
• Double slashes. /blog//spring-menu often returns the same page as the single-slash version. These usually come from a template joining a base URL ending in a slash with a path starting in one. Fix the template rather than adding more redirects.
• Uppercase and encoded characters. A slash rule that captures $uri in nginx uses the decoded path, while $request_uri keeps the original encoding. Use the form that does not change characters such as spaces or non-English letters in your URLs.
• Pagination and filters. Plugins sometimes build /shop/page/2 in one place and /shop/page/2/ in another. Include paginated and filtered pages in your sample.
• Anchors. The part after # is never sent to the server, so it cannot cause a duplicate and needs no rule.
• Old campaign links. Printed QR codes and ads may use the other form. The redirect covers them, which is one reason to redirect rather than return 404.

Keeping it consistent

Slash drift usually returns through new content rather than through server changes. A few habits stop it:

• Generate links from the CMS or framework's own link functions instead of typing addresses into templates by hand.
• Add a check to your release process that requests a handful of URLs in both forms and fails if both return 200.
• Keep the one-sentence URL rule in the same place as your redirect list, so whoever builds the next section of the site sees it.
• Run the Redirect Checker after any change of theme, plugin, framework version or hosting, because each can bring its own slash default.