A surprising amount of technical SEO is really server configuration. Whether a staging copy leaks into the index, whether www and the apex compete with each other, how fast a page comes back, and how long a new URL takes to get noticed are all decided below the application, and on most hosting you solve them with plugins that each do one of them badly.
Unicorn Panel puts them on one tab. This guide covers what is there and when to use it.
IndexNow: get new URLs noticed in minutes
Search engines traditionally find your changes by crawling. IndexNow inverts that: you tell them. It is a shared protocol that Microsoft Bing, Yandex, Seznam and a growing list of other engines read, and one submission fans out to every participating engine rather than needing separate registration with each.
Turn the toggle on under Websites > your site > SEO Tools > IndexNow and the panel does three things:
- Generates a random API key.
- Drops the
<key>.txtverification file into the site'spublic/folder, which is how the engines confirm you own the domain. - Schedules a daily job that submits only the URLs you have added, updated or deleted since the last run.
The delta matters. Submitting your whole sitemap daily is noise; submitting the eleven URLs that actually changed is a signal. Removed URLs are submitted too, so a page you unpublished gets flagged for de-indexing rather than being left to age out of the index on its own schedule.
The job appears on the Settings page under Cron Jobs with a Managed badge, and an IndexNow Log pane on the Logs page shows the result of every run, so you can see what was submitted rather than trusting that something happened.
Google is not in the IndexNow consortium. It keeps using its own sitemap and crawler path. So IndexNow is additive to a Google Search Console sitemap, not a replacement for it. Anyone telling you otherwise is selling something.
Staging mode: stop the copy outranking the original
The classic way to lose a week of traffic is to leave a staging copy of a site publicly reachable, let it get indexed, and end up with the staging domain competing against production for your own content.
The Staging mode toggle marks a site as a staging environment. The panel writes a proxy drop-in that stamps X-Robots-Tag: noindex, nofollow, nosnippet onto every response, so search engines that read the header leave it out of their index. A STAGING pill appears on the site's row in the Websites list while it is on, so you can see at a glance which sites are flagged.
Two things to be clear about:
It hides the site from search engines, not from people. It does not block visitors, rewrite links or change what the site serves. If you also need it hidden from casual visitors, layer Basic Authentication on top, which puts the whole site behind a username and password and lets you list IP addresses that bypass the prompt.
Turn it off before you point production traffic at the domain. A live site still stamping noindex is the same traffic loss in the other direction.
Canonical domain and HTTPS
Two toggles that cost nothing and are wrong on a depressing number of sites.
Force HTTPS 301-redirects every plain HTTP request to the HTTPS version of the same URL. Turn it on once the certificate has issued.
Preferred domain picks between www.yoursite.com and yoursite.com, and 301-redirects the other one to it. Without this, both resolve independently, and every link you earn is split between two URLs that search engines have to work out are the same page. Set it to neither only if you genuinely want both to resolve on their own.
Redirects
Panel-managed per-site redirects: a from-path, a to-URL and a status code of 301, 302, 307 or 308. They are written as single-location blocks at proxy reload rather than being processed by the application, so they cost nothing per request and work even if the application is down.
Use 301 for anything permanent. A 302 for a page you have actually moved for good is one of the most common ways to throw away the link equity you earned.
Alias domains
Serve the same site on several domains, each with its own Let's Encrypt certificate issued automatically. Useful for a rebrand where the old domain needs to keep working, or for country variants.
Worth pairing with a canonical strategy: several domains serving identical content with no preferred version is the same split-signal problem as www versus apex, at a larger scale.
Blocking crawlers you did not invite
Two separate denylists, because they are two different problems.
Block AI crawlers returns 403 to crawlers that identify as AI training agents, covering OpenAI, Anthropic, Google-Extended, Perplexity, Meta and a long tail of others, plus the data-collection frameworks that behave the same way. Regular search engines are unaffected, so turning it on does not cost you search traffic. Whether you want this is a business decision, not a technical one.
Block bad bots is a curated denylist of scrapers, site rippers, email harvesters and vulnerability scanners, updated with the panel. It also catches requests with an empty User-Agent, which is common for lazy scrapers.
Both match case-insensitively on a substring of the User-Agent, so a crawler advertising itself inside a longer Mozilla string is still caught.
Page caching, which is an SEO tool
Core Web Vitals are a ranking input, and the cheapest way to improve them is usually to stop generating the page at all.
The Website Cache serves repeat requests for the same page from memory without touching PHP. You can exclude specific pages, cookies and query strings, with * wildcards, so a cart or a logged-in session is never served from cache.
For WordPress, the Cache Helper mu-plugin purges automatically when you publish, edit or delete a post, change a theme or activate a plugin, and adds a Web Cache dropdown to the admin toolbar for manual purges. You can also purge specific URL patterns from the panel, or trigger one from outside it with an HTTP PURGE request.
Measuring it
The Analytics tab for each site samples the proxy's access log rather than relying on a JavaScript tag, so it counts visitors who block trackers. Visitors, pageviews, bandwidth, top pages, top referrers and country breakdown, with a compare mode that puts the current period against the one before it.
Log-based analytics and tag-based analytics measure different things and will not agree. That is expected: use the log numbers for traffic shape and the tag numbers for behavior.
A sensible order to do this in
- Force HTTPS and set the preferred domain. These are the ones that silently cost you the most.
- Staging mode on anything that is not production, before it gets indexed.
- Website Cache, with sensible exclusions.
- Redirects for anything you have moved.
- IndexNow, once the above is right. There is no point telling engines about URLs that then redirect or serve slowly.
- Crawler blocking, if you want it.
Frequently asked questions
Does IndexNow help with Google? Not directly: Google is not part of the consortium. It helps with Bing, Yandex, Seznam and the other participating engines. Keep your Search Console sitemap for Google.
Will staging mode remove a site that is already indexed? It stops it being indexed going forward and prompts engines to drop it over time, but if a staging copy is already indexed and ranking, you want to deal with that actively rather than wait.
Do the redirects cost performance? No. They are handled at the proxy layer rather than by the application.
Does blocking AI crawlers affect search rankings? No. The AI denylist and the search-engine crawlers are separate.
Are these on paid plans only? No. They are per-site toggles available at every tier, including the free one.
It should have come with the hosting
All of this is in the panel, per site, at every tier. Try the live demo or read the documentation.