Nothing we index goes unlabelled: every hostname and URL that passes through our pipeline carries at least one of the 59 labels set out on this page. The scheme groups online threats, age-restricted material, attention drains, commerce and money, the technical plumbing of the web, everyday interests and civic resources — shaped so a group can be pasted straight into a DNS resolver, a firewall rule set, a home parental control or a secure web gateway. Schools rely on the same scheme when filtering for K-12 schools across 59 content categories, and auditors use it to map our 59 categories to CIPA requirements.
The deny-by-default set. These are the labels most organisations and households switch off before touching anything else.
Domains our classification pipeline has marked as untrustworthy: payload delivery, credential-harvesting pages and other restricted material.
Paid VPN brands, open proxy front-ends and anything else designed to hide where traffic is really going. Workplaces and schools usually shut this group off.
Screen-sharing suites, remote desktop clients and unattended-support software. Useful to IT teams, and equally useful to anyone routing around the perimeter.
Redirect and link-shrinking services. Because the real destination stays hidden until the click lands, they show up constantly in restricted content campaigns.
Content built around hostility toward people for who they are, including supremacist and extremist publishing.
Gore, footage of violent acts and terrorism material, particularly pages that celebrate it rather than report on it.
Firearms and explosives across the board: retailers, trading boards, modification guides and instructional writing.
Storefronts, marketplaces and enthusiast pages covering controlled substances and the paraphernalia that surrounds them.
What a household filter, a school district or a CIPA audit will expect to find switched off.
Pornography and sexually explicit entertainment. This is the first label any classroom or parental-control policy disables.
Shops selling underwear, swimwear and intimate apparel — borderline enough that schools and stricter workplaces tend to include it.
Sportsbooks, casino platforms, poker rooms and state lotteries, a group that many countries regulate directly.
Distilleries, breweries, wine merchants and any site promoting or selling alcoholic drinks.
Cigarettes, cigars, vaping hardware, e-liquid vendors and the marketing built around them.
Matchmaking apps, dating portals and romance-oriented social platforms.
Rarely dangerous, frequently restricted between nine and five or during lesson time on office, campus and carrier networks.
Profile-driven networks where people follow each other and publish posts to a wider audience.
Game studios, digital storefronts, esports coverage, clan sites and titles playable straight in the browser.
On-demand and live audio and video: film catalogues, television, music services and user-uploaded clips.
Show directories, episode hosting platforms and the home pages of individual audio series.
Torrent trackers, peer-to-peer swarms and the client software that connects to them.
Consumer drives, personal backup providers and one-click file hosts, a common route for data leaving the network.
The everyday working web. On a corporate network these labels normally sit on the allow list.
B2B suppliers, consultancies, management advisory firms, trade press and general corporate web presences.
Investment houses, insurers, market-data vendors, brokerages and fintech products.
Retail banking portals where customers sign in and move money around.
Exchanges, wallets, DeFi front-ends, chain explorers and NFT trading venues.
Vacancy boards, applicant tracking portals, recruitment agencies and career guidance.
Ad exchanges, tracking networks, martech platforms and agencies that sell promotion.
Retail storefronts of every kind, from single-brand shops to catalogues at department-store scale.
Venues where many independent sellers list goods: auction houses, classified boards and open marketplaces.
Listing portals, letting agents, brokerages and property investment resources.
How people reach one another and put opinions in public. Policy here swings hard depending on the sector.
Webmail front-ends, chat apps and consumer instant-messaging services.
Threaded discussion boards, Q&A communities and single-subject bulletin boards.
Member-run platforms, neighbourhood groups and special-interest gatherings.
Individually written blogs, personal homepages and self-published commentary.
Newspapers, wire services, broadcasters, magazines and professional editorial outlets.
Party sites, election campaigns, lobbying groups and partisan commentary.
Broad and largely harmless subject matter: wellbeing, learning, food, travel, sport and hobbies.
Clinical references, patient information, wellness writing and healthcare provider sites.
Schools and universities, encyclopedias, dictionaries, study guides and reference libraries.
MOOCs, live training sessions, certification courses and professional skills platforms.
Peer-reviewed journals, research institutions and popular-science publishing.
Cinema, music, television, theatre, celebrity coverage and cultural criticism.
Flight and hotel booking, tour operators, national tourist boards and destination writing.
Menus, table booking services, delivery apps and restaurant review guides.
Recipe sites, cooking blogs, beverage guides and home-brewing or baking hobbyists.
League and club sites, fitness and training resources, outdoor pursuits and supporter forums.
Veterinary advice, breed information, shelters, rescue groups and wildlife conservation.
Carmakers, dealer groups, repair shops and parts suppliers, plus rail, bus and freight operators.
Renovation guides, interiors, landscaping, allotments and DIY project write-ups.
Congregations, scripture resources, faith communities and material on spiritual practice.
The layer everything else runs on, together with the press and the tooling that surround it.
Hardware and software coverage, IT how-tos, product reviews and engineering blogs.
Chat assistants, model playgrounds, agent frameworks and AI-driven productivity front-ends.
Repository hosts, build and CI services, editors and API documentation portals.
Browser-delivered SaaS, online utilities and productivity suites.
Download portals, open-source project pages and consumer application archives.
General search engines and other large-scale information retrieval services.
Edge networks, object storage, managed hosting and cloud platform endpoints.
Registered names serving nothing but a placeholder, an ad page or a sale listing.
State bodies, defence, and the fallback label that guarantees complete coverage.
Public agencies, statute and case-law databases, courts and regulatory authorities.
Armed services, defence manufacturers, service publications and veterans' organisations.
Where a legitimate site simply does not fit the other 58 labels. Holding this 59th category means no domain is ever handed back unlabelled.
Web Filtering Service ships the same taxonomy two ways: as a live REST API, or as an on-premise dataset of up to 100 million domains that you host yourself. Both routes return the 59 categories described above.
Browse Dataset Packages