Five Ways to Audit Links, From Manual Checks to Continuous Monitoring
An honest comparison of the main approaches to finding broken links and redirect problems, so you can match the method to the size of your site and the size of your team.
There is no single right way to audit links. A personal blog with two hundred pages and an online store with two million URLs have the same underlying problem and very different practical constraints. The approaches below sit on a spectrum from zero setup and low coverage to high setup and complete coverage, and most teams end up combining two of them: one for routine checks and one for the big passes around launches and migrations. We have rated each approach on the five criteria that decide whether an audit actually leads to fixes, and we have tried to be candid about where each one falls short, including the hosted approach that our own product belongs to.
| Option | Coverage of the full site | Setup and skill required | Speed on large sites | Ongoing cost and effort | Reporting and developer handoff |
|---|---|---|---|---|---|
| Manual spot checksBest for: Very small sites, a single page or section, or a quick sanity check after editing one article. | Very low. You check the pages you think of, which are rarely the pages that are broken. | None. Open the page, click the links, watch for errors. | Impractical beyond a few dozen pages. Redirect chains are invisible without developer tools. | Free in money, expensive in attention, and easy to stop doing. | Whatever you write down. Usually a message or a note with a handful of URLs. |
| Single-page link checkerBest for: Content editors who want to verify one page or one article before publishing, without touching the rest of the site. | One page at a time by design. It does not follow links to discover other pages. | Minimal. Typically a browser extension or a form that takes a URL and lists the links it found. | Fast per page, but there is no way to run it across a site without repeating the work manually. | Usually free or cheap. Effort scales linearly with the number of pages you care about. | A flat list of links and status codes for that page. No grouping, no chain maps, no history. |
| Desktop crawler you run yourselfBest for: SEO specialists and technical marketers who audit several sites, want full control over crawl settings, and are comfortable configuring software. | High when configured well. Follows links, reads sitemaps, and can render JavaScript, though rendering is slow on a laptop. | Moderate to high. You need to understand crawl scope, parameter handling, rate limits, and how to read the output. | Limited by your machine's memory and your connection. Very large sites may need to be crawled in segments or overnight. | Typically a license fee plus your own time for every run. Nothing happens unless someone remembers to start it. | Rich exports, but raw. Turning them into a prioritized task list is manual work every time. |
| On-demand hosted crawlBest for: Teams that need a complete audit of one site at a specific moment, such as before a launch or after a migration, without installing or configuring anything. | High. Runs from servers rather than a laptop, so large sites, deep crawls, and multiple URL sources are practical. | Low. Enter the domain, choose a few scope options, and wait for the report. The judgment about what to fix is still yours. | Good, and it does not tie up your machine. Crawl rate is usually throttled to protect the site being audited. | Pay per audit or per month. Cheaper than a specialist's hours for a one-off, less economical if you crawl daily. | Typically strong: grouped issues, redirect chain maps, and exports built for tickets, though quality varies by service. |
| Continuous monitoring with scheduled crawlsBest for: Large or fast-changing sites with a dedicated owner, where a broken link on a key page needs to be caught within hours rather than at the next audit. | High and repeated. Scheduled crawls catch regressions, and alerts on specific URLs cover the pages that matter most. | Moderate. Someone has to define scope, thresholds, alert rules, and who receives them, then keep those rules current. | Designed for it, but repeated full crawls of a huge site cost real money and may need to be segmented. | The highest recurring cost of the five, and alert fatigue is a genuine risk if thresholds are set carelessly. | Best for trend data and regressions. Can be noisy for the initial cleanup, where an on-demand audit is often clearer. |
- Manual spot checks: Useful as a final human check on a page that matters, never as the audit itself.
- Single-page link checker: A good editorial habit that catches outbound link rot early, but it tells you nothing about orphan pages or site-wide template issues.
- Desktop crawler you run yourself: The most flexible option and the one with the steepest learning curve; excellent in the hands of someone who runs audits regularly.
- On-demand hosted crawl: This is the category DeadLinkr belongs to, so weigh our view accordingly. It suits the audit-when-something-changes pattern better than constant surveillance.
- Continuous monitoring with scheduled crawls: Worth it once the backlog is clean and the site changes faster than a monthly audit can track; overkill before that.
Our verdict
For most small and mid-sized sites, the practical answer is a pairing. Use a single-page checker or manual spot check as an editorial habit for new content, and run a full crawl, whether desktop or hosted, at a fixed cadence and around every launch or migration. The full crawl is the one that finds orphan pages, template-level breakage, and redirect chains, which no page-at-a-time method can see. Pick desktop if you have a specialist who will run it regularly and enjoys configuring it. Pick hosted if you want the coverage without the setup, or if the person running the audit is not the person who will fix the results.
Continuous monitoring earns its cost on sites where a single broken landing page or checkout link costs more in an afternoon than the subscription costs in a year. Even then, start with a full on-demand audit to clear the backlog, because monitoring a site with thousands of existing issues produces noise rather than signal. Whatever you choose, the audit is only half the job. The other half is a grouped, prioritized export that a developer can act on, and a follow-up crawl that proves the fix worked.
Frequently asked questions
Do I need a full crawler if my site is small?
If the site has under a few hundred pages and changes rarely, a monthly automated crawl, hosted or desktop, which finishes in minutes on a site that size and doubles as the full audit, plus a page checker for new content is usually enough. Add a quarterly deeper look at redirects, PDFs, and external links. A full crawler still catches things spot checks miss, such as redirect chains created by hosting or CMS rules, so also run one around any structural change.
Can a hosted crawl audit a staging site that is not public?
Often, but it depends on the service and how staging is protected. Password-protected or IP-restricted staging environments typically need credentials or an allowlisted crawler address. If that is not possible, a desktop crawler run from inside the network is the fallback for pre-launch checks, followed by a hosted crawl of production after go-live.
Read the complete guide for the full reasoning behind this comparison.