How to set up URL redirects, site-wide SEO and the 404 page

Redirects send visitors and search engines from an old address to a new page, so old links and Google results keep working. This screen also holds the site's fallback SEO and the No index switch. Open the editor, click SEO/AEO, then Additional Settings.

How to add a URL redirect

1. In the URL Redirects card, enter the Old path (for example /old-page).
2. In Redirect to, type a page of this site (suggestions appear as you type) or a full https:// address.
3. Choose the Type: Permanent (301) for a page that has moved for good, or Temporary (302).
4. Click Add redirect.

You'll see "Redirect added — it works on the live site now". Redirects work on the live site straight away, with no republish. The table shows how many visits each redirect has had on the live site (previews aren't counted).

Redirects added for you. When you change a published page's address with ⋯ → Page URL in the Pages panel, a 301 redirect from the old address to the new one is added here automatically (one per entry for a dynamic page). See Pages panel. A page that's still at an address always wins over a redirect from it, so the old address keeps showing the page until you publish the site with the new address.

Managing the list

• Search: type part of an old or new address in the search box above the table.
• Edit: click the pencil on a row, change it, then click Save.
• Delete: click the bin, then Delete to confirm. To delete several, tick their boxes (or the box in the header for every row shown) and click Delete selected, then Delete.
• Export CSV downloads every redirect with the columns Old path, New path, Type.
• Import CSV: paste one redirect per line, like /old-path, /new-path, 301. An exported file can be pasted straight back in. A header row such as Old path, New path is skipped. The result tells you how many were added and how many were skipped.

How redirects behave:

• Capital letters and a slash at the end don't matter. /Old-Page/ and /old-page are the same old address. An old address with a slash at the end, like /about-us/, goes straight to its new page in one step.
• Chains are shortened. If you point a redirect at an old address that already redirects somewhere, it's saved pointing straight at the final page, and a note says so (for example "Pointed straight to /new-page, because /old-target already redirects there"). Redirects that pointed at an address you've just redirected are updated the same way. A temporary (302) redirect in the middle is left alone, because it may change back.
• Tracking stays on. Anything after ? in the visitor's address, like utm_source or gclid, is passed on to the new page.
• A page wins. A redirect only works for an address with no page. If a page still uses the old address, delete or rename the page first.
• A dynamic page's own address (like /services, when items live at /services/roof-repair) goes to the home page or a page you choose (see Collections and dynamic pages). A redirect you add for that address wins over both.
• No loops. You can't add a redirect that would send visitors round in a circle, like /a → /b while /b → /a.
• Search engines see them as 301 and 302. Technically, permanent redirects are sent as 308 and temporary ones as 307. Google and Bing treat these exactly like 301 and 302.
• Images can't be redirected: addresses ending in .jpg, .jpeg, .png, .gif, .webp or .svg, or starting with /images/ or /templates/. Pages and PDFs can.
• Redirects work on the site's live address only after the site has been published.

When you've moved a client from an old website, see Redirects for a new site to match the old pages to the new ones.

Site-wide SEO

The Site SEO card holds the fallbacks:

• Site meta title and Site meta description are used for any page without its own. A page's own SEO settings always win.
• Site keywords (optional): major search engines ignore these.
• No index for the entire site tells search engines not to show any page. Use it for staging or test sites, and remember to turn it off before launch. While it's on:
  • robots.txt blocks every crawler (Disallow: /), including AI crawlers;
  • sitemap.xml and llms.txt stop being served (they return "not found");
  • each page carries a "noindex" tag once you republish.

Click Save, then republish the site. The robots.txt, sitemap and llms.txt changes don't wait for the republish, but can take up to an hour to show (they're cached). See Sitemap and robots.txt.

The 404 page

Visitors see the 404 page when they open an address that doesn't exist. It uses your header and footer and tells search engines the page isn't there. To change it, go to Settings → 404 page and click Customize. The page opens in the editor. Revert to original puts back the built-in page. A redirect for an address always wins over the 404 page. Publish the site to put 404 changes live.

Examples

• Renamed a page: /our-services became /services. Add /our-services → /services, Permanent (301).
• A seasonal offer: send /offer to /summer-sale as Temporary (302), then point it at /winter-sale later.
• A move from an old website: export the old site's page list, build lines like /old-page, /new-page, 301, and paste them into Import CSV.
• Staging copy: turn on No index for the entire site on a test copy so Google never lists it.

Limits and what it can't do

• One redirect per old address.
• Images (.jpg, .jpeg, .png, .gif, .webp, .svg, and anything under /images/ or /templates/) can't be redirected.
• A redirect can't override a page that exists at that address.
• No wildcard or pattern redirects: each old address is its own row.
• Visit counts only include the live site.
• Redirects aren't saved in site backups, so restoring a backup leaves them as they are now.

Who can use it

• Agency people with the Mount Olympus tool.
• Client people need SEO settings to add, edit, import or delete redirects and change site SEO (people with Site Audit can view the list). The 404 page is in Settings and needs Editor.

Troubleshooting

• The redirect doesn't work: check that no page still uses the old address (a page wins), the site is published, and you're testing the live address, not a preview.
• "Enter the old page path, like /old-page": the Old path is empty or just /. The home page can't be redirected.
• "Redirect to a page path (/new-page) or a full web address (https://…)": the Redirect to box is empty or malformed.
• "A page can't redirect to itself": the old and new addresses are the same.
• "That would send visitors in a loop — another redirect already points back here": remove or change the other redirect first.
• "That path already has a redirect": edit the existing row instead of adding a new one.
• Import says some were skipped: those lines were blank, malformed, duplicates or would loop. Fix them and paste just those lines again.
• Broken links in the Site Audit: point the link at a page that exists, or add a redirect here.