Robots.txt first
Before anything else we read robots.txt and obey it, Crawl-delay included. A Disallow for TellstackBot means we fetch nothing more and only read public DNS.
Spot the
stack
bbc.co.uk
- Drupal 10meta generator: Drupal 10 (https://www.drupal.org)
- Varnish header via: 1.1 varnish
- PHP implied by Drupal
- RequireJS script: require.min.js
- Slack DNS TXT: slack-domain-verification
- HSTS header strict-transport-security
An API that names the technology behind any website, with versions and the evidence. From £49 a month, not £250.
Get a free keySightings from our crawl
Real sites from the Tranco top list, as our crawler found them. Choose one to read its markings.
Read the markings
- Name and category7,647 technologies, from CMSs and shops to payment, chat and A/B testing.
- VersionWhere the site shows one: a generator tag, a script path, a header.
- ConfidenceEvery independent tell adds to it, up to 100. Filter lists by it.
GET /v1/lookup?domain=blog.example
{
"domain": "blog.example",
"status": "ok",
"cached": false,
"scannedAt": "2026-10-28T08:41:12Z",
"technologies": [{
"name": "WordPress",
"categories": ["CMS", "Blogs"],
"version": "6.5.2",
"confidence": 100,
"evidence": [{
"type": "meta",
"key": "generator",
"match": "WordPress 6.5.2"
}]
}, {
"name": "PHP",
"confidence": 100,
"evidence": [{
"type": "implied",
"key": "WordPress"
}]
}]
}- EvidenceThe header, cookie, script, meta tag or DNS record that gave it away.
- StatusRead, blocked us, robots.txt says no, or unreachable. Never a guess.
- FreshnessFrom the index if scanned in the last 30 days; live in seconds if not.
How we watch
The way a careful birder works: from the path, quietly, and never by flushing anything out of the hedge. Each site gets a handful of requests a month, and a site that asks us to leave is left alone.
Three pages, slowly
The homepage and at most two pages it links to, with at least a second and a half between requests to one host. Our user agent names us and links to a page explaining how to opt out.
Blocked means stop
A bot check, a challenge page or a 429 ends the visit. No proxies, no disguises, no retries for 90 days. The result says “blocked”, plus whatever public DNS tells.
Every tell counted
Headers, cookies, scripts, meta tags, page markup, DNS records and the TLS certificate run through 7,647 open fingerprints. Each finding keeps the evidence.
Changes logged
Each rescan is compared with the last good one. Added, removed and upgraded technologies go into the site’s history and out to your webhooks.
Pricing
Free
For trying it on the sites you care about.
100 lookups a month, 10 watches, the top 25 rows of any list, the whole API. No card.
Pro
For sales teams, agencies and anyone enriching a CRM.
10,000 credits a month per block: a credit is one lookup or ten rows of a list. 500 watches, full lists and CSV exports, webhooks. £41 a month billed yearly; £39 a block from five blocks. Cancel any month.
£49/month
Get ProQuestions
How is this different from BuiltWith or Wappalyzer?
Same job, priced for people who aren’t enterprises. BuiltWith starts at $295 a month and Wappalyzer at $250. Tellstack starts free, and 10,000 lookups cost £49. You also get the evidence behind every finding, not just a name.
Where do the fingerprints come from?
From the open Wappalyzer fingerprints, which the community keeps up to date as webappanalyzer: 7,647 technologies across 109 categories. We add DNS and certificate checks and run them on our own servers.
What can’t it see?
We read pages the way a plain browser fetch does, without running JavaScript. Technologies that only show up once scripts run, mostly some single-page app frameworks, are missed. Most CMSs, shops, analytics, CDNs and hosts show up without it.
What about sites that block bots?
We stop. A site that challenges us gets status “blocked” and is left alone for 90 days. You still get what its public DNS gives away, such as its email and DNS providers.
Can I get lists of sites using a technology?
Yes, from our own index, most visited sites first, filtered by country domain and by the date we first saw it. The index grows from the Tranco top list and from every lookup. It is young: say what you need and we’ll crawl that corner first.
How fresh is it?
The index is rescanned monthly and watched sites weekly. Any lookup can ask for a live scan, which takes a few seconds.