If you sell to companies based on what software they run ("find me every site on Shopify", "which agencies' clients still use Universal Analytics"), you have three ways to get that data in 2026. They differ in price by a factor of about 100, and most people only know the expensive two.
Option 1: BuiltWith
The heavyweight. Its real product is historical data: what a site ran three years ago, when it switched, technology trends across the whole web. If you need that, nothing else comes close. Plans commonly run hundreds of dollars per month, and for trend analysis it earns it.
But most sales and SEO use cases don't need history. They need "what does this list of 2,000 domains run today", and paying a monthly subscription for that is renting a combine harvester to mow a lawn.
Option 2: Wappalyzer
The origin of most detection tech. The browser extension is great for spot checks on a single site. The API and bulk products moved to paid plans years ago, and the open-source fingerprint dataset the community built lives on in maintained forks.
Option 3: pay-per-result detection
The newer model: no subscription, you pay per site checked, typically around half a cent. The engines run the same class of open-source fingerprints (7,500+ technologies: CMS, ecommerce, analytics, frameworks, CDNs, with versions and confidence scores), so detection quality on current-state checks is comparable. What you give up is history; what you get is a bill that matches your actual usage.
Worked example with real numbers: a 1,000-domain prospect list costs about $5 to enrich with full tech stacks at pay-per-result pricing. The same job through incumbent per-result pricing on marketplace tools ranges from $70 to $100, and through a subscription product it costs whatever your monthly plan costs whether you run the list or not.
When each makes sense
Historical trends and web-wide statistics: BuiltWith, no contest. One-off manual checks while browsing: the Wappalyzer extension, free. Bulk current-state enrichment for lead-gen, SEO audits, or competitive research: pay-per-result is the price-performance winner, and it fits AI agents too, since MCP-compatible tools let an agent enrich a list mid-workflow and the bill lands in cents.
Full disclosure: I built one of the pay-per-result options, Website Technology Detector (about $5 per 1,000 sites, MCP-ready, live example output on the listing). If your use case is the third one, it will save you real money; if it's the first, buy BuiltWith and don't look back.
Top comments (0)