In Part 4 we went through three Google Search Console exports and ten directory listings for InterviewAce. This part is more practical: the script we use to check a listing, and what it found when we ran it on twelve more.

The full write-up of those twelve, with each directory's catch, is on our other blog: 12 SaaS Directory Listings, Re-Checked (Part 4). This post is the tooling behind it.

Why a Script

Opening a listing in a browser and glancing at the "Visit Website" button tells you almost nothing. A link can look fine and still be useless to a search engine because of any one of these:

  • rel="nofollow", sponsored or ugc on the anchor
  • noindex or nofollow in the page's robots meta tag, or in an X-Robots-Tag header
  • a canonical tag pointing at some other URL
  • the path being disallowed in the site's robots.txt
  • your domain mentioned in the text but never actually linked

None of that is visible without reading the HTML, and with dozens of listings, reading HTML by hand stops being realistic.

The Script

Node 18 or later, no dependencies. Save it as audit-listing.mjs, since top-level await needs an ES module.

// audit-listing.mjs — check whether a directory listing really links to your site // Usage: node audit-listing.mjs yoursite.com https://directory.example/item/yourproduct const [, , site, url] = process.argv; const UA = 'Mozilla/5.0 (compatible; listing-audit/1.0)'; const res = await fetch(url, { redirect: 'follow', headers: { 'user-agent': UA } }); const html = await res.text(); const u = new URL(res.url); // 1. every <a> pointing at your site, with its rel attribute const links = (html.match(/<a\b[^>]*>/gi) || []) .filter(a => new RegExp(`href=["']https?://(www\\.)?${site.replace(/\./g, '\\.')}`, 'i').test(a)) .map(a => (a.match(/rel=["']([^"']*)["']/i) || [])[1] || '(none)'); const followed = links.filter(r => !/nofollow|sponsored|ugc/i.test(r)); // 2. page-level robots, from the meta tag and the response header const meta = [...html.matchAll(/<meta[^>]+name=["']robots["'][^>]*content=["']([^"']+)/gi)].map(m => m[1]).join(', '); const robots = (meta + ' ' + (res.headers.get('x-robots-tag') || '')).toLowerCase(); // 3. canonical const canonical = ((html.match(/<link[^>]+rel=["']canonical["'][^>]*>/i) || [''])[0].match(/href=["']([^"']+)/i) || [])[1] || ''; // 4. robots.txt (rough check: Disallow prefixes in the User-agent: * group) const txt = await fetch(u.origin + '/robots.txt').then(r => (r.ok ? r.text() : '')).catch(() => ''); const group = (txt.split(/user-agent:\s*\*/i)[1] || '').split(/user-agent:/i)[0]; const blocked = [...group.matchAll(/^\s*disallow:\s*(\S+)/gim)].map(m => m[1]).some(p => u.pathname.startsWith(p)); console.log({ status: res.status, linksToYou: links, dofollow: followed.length > 0, noindex: robots.includes('noindex'), pageNofollow: robots.includes('nofollow'), canonicalElsewhere: !!canonical && canonical.replace(/\/$/, '') !== res.url.replace(/\/$/, ''), blockedByRobotsTxt: blocked, mentionedButNotLinked: links.length === 0 && html.toLowerCase().includes(site.toLowerCase()), }); 

Run it like this:

node audit-listing.mjs interviewace.online https://dayslaunch.com/item/interviewace 

What it deliberately does not do

  • It does not run JavaScript. It reads the HTML the server sends. If a directory builds its links in the browser, the script reports no link. That is useful in itself: a crawler sees the same raw HTML first.
  • The robots.txt check is rough. It only looks at Disallow lines under User-agent: *, and ignores Allow overrides and wildcards. Good enough to flag a problem, not to settle one.
  • It cannot tell you whether Google has found the page. For that, compare against Search Console's Links report across several exports.

What It Found on 12 Listings

All twelve passed on November 22, 2026: status 200, at least one followed link to our site, no noindex, no page-level nofollow, canonical pointing at the page itself, and not blocked in robots.txt.

Directory Date shown on the page What the script showed
LaunchIt Launched Sep 5, 2026 Two followed links
Findly.tools Listed Sep 11, 2026 One followed link
LaunchNest none One followed link, two nofollow
Days Launch Published Aug 27, 2026 Three links, no rel at all
The One Startup Published Aug 27, 2026 Three links, no rel at all
DeepLaunch.io Listed Jul 23, 2026 One link, no rel
Turbo0 Published Aug 9, 2026 Three followed links
Roozna none One followed link; visible text rendered by JavaScript
NavFolders Published Nov 14, 2026 Three links, no rel
ShipThing none Two followed links
Looking For Tools none Two followed links
Good for Bots Scanned Nov 8, 2026 One followed link

A few things worth pointing out.

Mixed links on one page are normal. LaunchNest has one followed "Visit Site" link and two nofollow links to the same domain. If you only checked the first link you found, you could reach either conclusion.

No rel at all is the cleanest result. Days Launch, The One Startup, DeepLaunch and NavFolders put no rel on the anchor. A rel with only noopener or noreferrer is fine too. Those affect the browser, not ranking.

Raw HTML and the rendered page can differ. Roozna's link is in the HTML, but almost all of its visible text is rendered by JavaScript. Search engines can render it, usually later than they read the HTML.

Passing the script is not the same as being useful. Turbo0 passes, but it is a directory for content creators. ShipThing and Looking For Tools pass, but they run on the same template. Good for Bots passes, but it is a report about our site, not a listing. The script answers the technical question; judgement answers the rest. The justblogged write-up goes through those calls one by one.

The Surprise: Our Own Posts

Out of curiosity, we pointed the script at our own dev.to posts:

{ dofollow: true, noindex: true, pageNofollow: true } 

Every link in the post is a normal followed link, but the page itself carries <meta name="robots" content="noindex"> and nofollow. Most of our dev.to posts look like this; at least one does not.

It is the lesson this series keeps coming back to: check the page, not just the link. A perfectly clean anchor on a nofollow page passes nothing. If you write about your own product anywhere, run the same check on those pages too.

Where We Keep the Results

Listings that pass go on our own site, so each one has at least one crawlable path pointing at it:

The rest of the series