Hoppla crawls your site, follows every internal link, and writes down the ones that don't answer: the HTTP status or error each returned and the page it's sitting on. You decide how deep it goes and how hard it knocks. It runs in the background and reports back when the run finishes. This one's advisory: a clean list of what's broken and where, not a 3am downtime page.
crawl acme.io · 487 pages · 6 broken (3×404 2×timeout 1×500) · 4m12s
Fictional values. Not a live measurement.
Hoppla starts at your URL and crawls outward, following the links on your own site page by page and testing each one for reachability. A link that doesn't answer gets recorded with the exact HTTP status it returned (a 404, a 500) or the error (a timeout) and the page it was found on. You set how far the crawl goes and how hard it hits. The run happens in the background, and the report lands when it's done.
Hoppla starts at your site and walks the link graph, discovering pages as it goes and collecting every link on your own host along the way. You cap it with max pages (10–5,000, default 500) and max depth (1–10, default 10), so it covers what matters without crawling forever.
Every link it finds gets requested for reachability. One that fails is logged with the HTTP status it returned and the page it sits on, so you get '404 on /pricing', not just a count. Links to other hosts, including other subdomains, are skipped, so the crawl stays on your site.
The check returns immediately and the crawl runs async. Request rate is yours, 0.5 to 10 req/s, default 2, so a fragile origin doesn't get hammered. A 30-minute deadline guards every run. A crawl that hits it stops and reports what it found on the pages it reached.
When the run completes, the broken links arrive as a list: each unreachable URL, its status code, and where it was found. This is advisory. A content-quality report you act on during the day, never a downtime incident that pages your on-call.
A spot-check of your top pages misses the dead link buried three clicks deep in a blog post from 2022. Hoppla crawls up to 5,000 pages and 10 levels down, so the link rot you'd never click your way to still surfaces.
A number that says '6 broken links' is a chore. Hoppla records each one with its HTTP status and the page it's on. A 404 on a missing redirect target and a flaky 500 from a backend route are two different fixes, and you can tell which is which on sight.
Page count, depth, and request rate are all dials you set. Throttle down to 0.5 req/s for a delicate origin, or open it up to 10 for a static site that can take it.
Your docs are your support team. A dead link in the install guide quietly drops readers on a 404, right when they're trying to get started. You're never going to click through every doc page yourself. So point a crawl at the docs, set it loose in the background, and get back the exact pages where links rotted, each with the status code that broke them. Fix the three that matter, skip the noise.
You just migrated a client's site and the URLs all changed. Half the internal links and every old anchor now land somewhere they shouldn't. Crawl the new site once, hand the client a clean report (here are the 14 links the migration broke, here's what each one returns) and fix them before launch instead of after a stakeholder finds the first 404. That's the deliverable that turns a risky migration into a tidy one.
Across hundreds of content-heavy client sites, link rot is constant and it's nobody's job. Schedule the crawl per site and let it run in the background at a request rate that won't disturb production. The report comes back with each broken link and its status code, ready to drop into the monthly report the client already pays for. So an account manager can say 'we caught and fixed 23 dead links this month' instead of waiting for the client to report the first one.
A scheduled link audit is a content-quality tier you can sell on top of hosting, fully under your own brand and domain. Each customer gets a recurring crawl with a branded report of what's broken and where, no Hoppla name anywhere in it. And because you set the request rate per crawl, even a customer on shared infrastructure gets audited at a rate their site can take.
Import your Oh Dear monitors into Hoppla. Then run both side by side for 24 hours or 7 days before you switch.
The importer opens in your workspace once you have access.
Get early accessThe crawl-and-report mechanics are similar. The difference is that in Hoppla the link check isn't a separate tool you log into on its own. It runs on the same monitor as your uptime, SSL, and DNS checks, with one bill and one alert config. And you control the crawl precisely: page count, depth, and request rate are all dials, so it scales up to a 5,000-page site or backs off for a fragile one, instead of running one fixed way.
Only as hard as you let it. Request rate is yours to set from 0.5 to 10 req/s, and it defaults to a gentle 2. Together with the max-pages cap, you decide exactly how much load the crawl puts on your origin: drop it low for delicate infrastructure, open it up for a static site that can take it.
No. The broken link check is advisory. It produces a content-quality report, not a downtime incident, so it won't fire your outage alerts or post to your status page. Link rot is something you schedule and handle during working hours, not an emergency that wakes you up.
Because the crawl runs async in the background. The check returns immediately so it isn't blocking anything, and the crawler keeps working through your pages. When it's done, the broken links land as a list with their status codes. Every run is guarded by a 30-minute deadline, and one that hits it stops and reports what it found on the pages it reached.
Checks, alerts and status pages in one workspace. Now in early access: join the list for an invite.