Collection
Plenty of the classic web assumes a mouse and a wide window. Everything in this collection holds up on mobile.
835 sites · page 10 of 18
The open music encyclopedia much of the industry quietly relies on, editable by anybody and free of any label's marketing department.
Long beatless mixes chosen for the state of mind where code actually goes in, delivered as plain numbered episodes.
Noughties television as it actually was, which is more unsettling than the eighties version because you remember most of it.
A simulated television set showing nothing but 1980s programming and adverts, complete with channel dial and static between stations.
The same idea moved a decade forward. Flip channels through the nineties and land somewhere between a cartoon and a cereal advert.
The default anime and manga database and tracker since 2004, and the source of the scores everyone else ends up quoting.
The largest soundboard on the web with a search box, which makes it either a toy or a workplace hazard.
Hundreds of calibrated soundscapes with ten-slider equalisers, built by an audio engineer who takes tinnitus relief seriously.
Fictional US national park built inside a colossal living organism, documented entirely as tourist brochures and incident reports.
Type a surname and get where it came from, what it originally meant and where in the world it still clusters.
The American government's public database of missing and unidentified persons, searchable by absolutely anyone.
Every photograph, video and audio recording NASA has released, searchable and free to use for anything.
The official live stream, which for long stretches is simply Earth turning below the space station.
Canada put its entire animation and documentary archive online for free, and some of it is the strangest work any government ever funded.
The official Atlantic and Pacific hurricane advisories, cones of uncertainty and discussion notes written in plain prose by duty forecasters.
Run a country by answering one absurd policy dilemma a day, then watch your statistics drift somewhere you did not intend.
The live feed from a deep-sea research ship's remote submersible, complete with scientists audibly losing composure over a new sea cucumber.
Neal Agarwal's collection of interactive toys and explainers, every one of them built far better than it needed to be.
A blog of neat things — odd history, strange products, good comics — running since 2005 with no algorithm involved.
A free host built as a deliberate successor to GeoCities, full of hand-written personal pages by people who wanted to make one.
Wordle for equations: eight tiles holding one arithmetic statement that has to balance as well as being correct.
Rhizome's restored survey of a hundred works of net art, each with the context needed to understand why it mattered.
Board games rebuilt for a group on a video call, with each person playing from their own phone. Codenames and Wavelength included.
The link at the end of every bait-and-switch since 2007, and the reason nobody fully trusts a shortened URL.
UK spoof news updated several times a day, best read as running commentary on whatever is currently going wrong.
A morning briefing of good stories, written for people who want a news habit that does not ruin breakfast.
Thirty years of usability research written up plainly and still free to read, which is why everyone cites it.
A permanent vaporwave station with fake adverts, a virtual room and album art on a CRT, running continuously for about a decade.
A ninja, flexing, to a backing track, at a frame rate suggesting nobody involved was trying especially hard.
The thirty-minute aurora forecast as an oval of probability drawn over the poles, which is the only honest way to plan a night out for it.
Rain, wind, fan and coffee-shop loops you layer into one background mix, then save as a preset for the rest of the week.
Retail and service workers' stories about customers, filed daily, and a bottomless argument for tipping generously.
Fast, minimal tracker that pings you the moment a followed episode airs. Open source and deliberately kept lightweight.
One well-sourced strange story per weekday, told at exactly the length of a cup of coffee.
London online radio with thousands of shows in the archive, most presented by somebody with an unreasonable record collection.
The pop-tart cat, the rainbow and the song, with a counter for how many seconds of your life you have handed over.
Nine hundred thousand digitised items from the New York Public Library, a large share of it public domain.
Tagged great whites ping their position whenever a fin breaks the surface, each shark named, weighed and given its own profile page.
Feed of products that should not exist, from bacon-scented soap to a hammock designed to hang under your desk.
The LiveJournal community that outlived LiveJournal, still posting celebrity news with the comment threads doing the work.
Eldritch horror podcast set in an alternate Appalachia where the mountains are older than people and considerably less friendly.
Search hundreds of thousands of scanned historical maps by place and date, held by libraries worldwide.
Radio drama, comedy and serials from before television took over, in quantities that would take several lifetimes to finish.
An encyclopedia of the world's writing systems, including the invented, the extinct and the barely attested.
A million checkboxes shared by everyone at once. Ticking them was pointless right up until people started writing messages in binary.
One minute of footage from a park somewhere in the world, matched to the exact minute you happen to be watching.
Screenshots of archived GeoCities pages posted automatically, one after another, as a rolling portrait of the amateur web.
Where English words came from, sourced and dated, compiled by one man who has been at it since 2000.