Smart 404 Redirect Suggestions Powered by Real Traffic

Every 404 error on your site is a missed opportunity. A broken page that once ranked in search results is now sending visitors nowhere. A URL someone bookmarked, linked to, or shared returns an error. The link equity it built up over years evaporates, hop by hop.

The hard part has never been fixing a 404 once you know about it. Creating a redirect rule takes seconds. The hard part is finding them — and more importantly, knowing which ones actually matter.

Today we're shipping smart redirect suggestions: an automatic pipeline that discovers your broken pages, enriches them with real traffic data, and surfaces the ones worth acting on first.

How redirect suggestions work

When a visitor hits a 404 on your site, that event doesn't have to disappear into a server log you'll never read. With the edge proxy enabled, we capture every broken page hit in real time — path, status, and the type of traffic that triggered it. That data feeds directly into your suggestions list.

But a raw list of broken paths is only the start. Not every 404 is created equal. Some broken pages are getting hammered by automated scanners that are completely irrelevant to your SEO. Others are receiving real organic search traffic every week — meaning Google is actively sending people to a page that no longer exists.

The difference between those two scenarios is everything when you're deciding where to spend your time. That's why we built traffic type classification into suggestions from day one.

Traffic type classification: search, human, LLM, and bot

Every suggestion in your dashboard is now tagged with the type of traffic that hit it:

Human — a real visitor typed or clicked their way to that URL
Search — an organic search crawler (Googlebot, Bingbot) crawled the broken path
LLM — an AI crawler (GPTBot, ClaudeBot, Gemini) accessed the broken path
Scanner — an automated tool like Screaming Frog, Ahrefs, or a security scanner hit it

A 404 that's only getting scanner traffic is noise. A 404 that's getting human visitors and search bot crawls is a problem you should fix this week.

The smart spam filter we built on top of this classification automatically deprioritizes suggestions that are bot-heavy with no human or search signal. What you see in the dashboard is already filtered — the list reflects pages that real users and search engines care about.

Google Search Console integration: find what's actually costing you rankings

The traffic signal from your edge proxy tells you what's broken *right now*. Google Search Console tells you what was broken *before you even noticed*.

GSC sits on top of months of historical data about which URLs Google has crawled, which ones are returning errors, and which ones were driving organic clicks before they went dark. When you connect your GSC account, we pull that data and layer it on top of your suggestions.

The result: you can see exactly which broken pages have active search impressions — meaning Google is still trying to send traffic to them. A page that gets 500 monthly impressions and is returning 404 is a redirect you should have created six months ago. GSC helps you find those pages before another month of rankings slips away.

Connecting GSC takes about a minute from your domain settings. Once connected, your next suggestions sync picks up the enrichment automatically.

Bulk imports from your SEO tools

Not everyone has the edge proxy enabled yet, and even those who do sometimes want to cross-reference data from tools they already use. That's why we built support for bulk CSV imports from:

Ahrefs — broken backlink exports, crawl errors
SEMrush — site audit broken page reports
Screaming Frog — full crawl exports filtered to 4xx status

Upload the file and your suggestions list is immediately populated with the broken pages your SEO tools found. The same traffic classification and spam filtering applies — if Screaming Frog's own crawler is the only thing that hit a path, we'll flag it accordingly.

Every source flows into the same suggestions view. You get one unified list across proxy traffic, GSC data, and whatever crawl tools you're already running — not three separate reports to reconcile.

What's coming next

The suggestions pipeline we shipped today is the foundation for something more ambitious: automatic 404 healing.

Sitemap scanning for every domain

Right now, suggestions are richest for domains with the edge proxy enabled. We're building a sitemap diffing system that extends this to any domain — with or without the proxy.

The idea is straightforward: we periodically fetch your sitemap, compare it against the previous version, and HTTP-probe any URLs that disappeared. If a page drops off your sitemap and returns a 404, we flag it as a suggestion. You don't have to do anything. We'll scan it automatically.

This means every site, regardless of setup, gets broken page detection out of the box. And because we store snapshots of your sitemap over time, you can see the exact moment a page went missing.

Automatic redirect creation

We're close to shipping full auto-heal: when we find a broken page with high confidence about the right redirect destination — based on path similarity, page title matching, and recency signals — we'll write the redirect rule automatically.

You'll still have control. Rules flagged as "review" confidence come to you for approval. Rules with high confidence scores get created immediately, with a full audit trail showing why. For chains (A → B → C), we flatten them to a single hop at the edge. No more redirect chains accumulating over time.

Traffic composition dashboard

We're building a richer analytics view that breaks down all your redirect traffic by type — not just total hit count, but how much of it is human, how much is search, and how much is automated. For proxy-enabled domains, you'll also see the breakdown on proxy-broken events.

This gives you a clear picture of what your site's traffic actually looks like — and surfaces domains that are getting hammered by bots so you can decide whether to investigate.

Redirect chain detection

Redirect chains silently degrade your SEO. When rule A sends traffic to rule B which sends it to rule C, every hop costs you PageRank signal. We're building chain detection directly into the redirect rules table — inline warnings on rules that form chains, with one-click resolution.

The bigger picture

We're building toward the most complete redirect management platform on the web — and we think about that goal through the lens of the edge.

When your traffic runs through our edge network, we see everything: every request, every broken URL, every bot hit, every redirect chain. That visibility is what makes the features above possible. We can detect problems before you notice them, classify traffic automatically, and write redirect rules without human intervention.

The edge also opens up capabilities that don't exist in any other redirect tool:

Bot traffic filtering at the edge — bot traffic never reaches your origin server. Your hosting costs stay down, your analytics stay clean, and your real traffic is easier to reason about.
User-agent routing — serve different experiences to different visitors based on who they are. Search bots can see canonical content; LLM crawlers can see structured data; human visitors get your full experience.
Redirect chains flattened automatically — we resolve chains at the point of request, so your visitors always reach the final destination in a single hop.

This is what makes redirect management at the edge different from managing .htaccess files or adding rules to your CMS. The power comes from operating at the request layer — before your origin server ever sees the traffic.

Simple to use. Clear pricing.

We're building toward the most complete redirect platform on the web, but our goal has never been to build the most complex one. Every feature we ship has to be understandable in a few seconds and actionable in under a minute.

The pricing works the same way. You pay for the requests you run through the system. No seats, no feature tiers, no hidden limits on the things that matter.

If you're not yet using redirect suggestions, connect your domain and enable the edge proxy to start seeing your 404s in real time. GSC connection is optional but recommended — it's the fastest way to find which broken pages are actively costing you search rankings.

If there's a data source or suggestion type we haven't covered yet, we want to hear about it. The pipeline is designed to add new sources easily — if your SEO tool exports a broken page report in a format we don't support yet, tell us and we'll add it.