How we verify
The claim is that everything here has been checked. This is what that means.
Most event sites are measured by how much they have. This one is measured by how much it threw away — because the expensive thing is not the ticket, it is the evening.
We have gathered 54,631 listings from 38 sources. 439 of them are on this site. 312 were rejected outright; the rest were duplicates of something already here, or turned out not to be events at all.
These numbers are read from the database as this page loads. They will be different tomorrow.
The six gates
A listing has to pass all six. Failing any one sends it to a review queue or quarantine — never to the site.
01
Relevant
A classifier decides whether this is genuinely professional networking, and has to write down why. Webinars in disguise, recruitment pitches and sales seminars are the three we reject most. Anything the classifier is unsure about goes to a person, not through.
02
Complete
A title, a start time with a real timezone, a working link, an organiser, a description, and a definite answer on four things: is it online or in person, what does it cost, where did we get it, and which town is it in. A blank we have quietly defaulted counts as a failure, not an answer.
03
Fresh
Listings are re-checked as the date approaches — daily inside a week, twice daily inside 48 hours. Cancelled events are marked. A link that stops working takes the listing off the site rather than waiting for you to find it.
04
Clean
Titles are normalised — no SHOUTING, no emoji, no date stuffed into the name. Descriptions are rewritten from the facts and never copied. Where an organiser publishes their own event image we display it from their server; we never copy it onto ours.
05
Deduplicated
The same event listed on Eventbrite, Meetup and the organiser's own site is one entry here, with each place you could book it attached to it. You should not have to work out that three results are one evening.
06
Honest
Every listing carries the date we last checked it and a way to tell us it is wrong. Organiser corrections and removals are honoured within 48 hours — argued with never.
Where we bend, and why
Four exceptions, written down. A standard with unlisted exceptions is not a standard — and we learned that the hard way: an internal audit once scored this index at 48% because the biggest exception was the one nobody had documented. The true figure was 100%.
- Some organisers only release the address after you register. Those publish without a venue and say so, rather than showing you a city centre pin that means nothing.
- A venue we cannot place on a map publishes with its name and address but no pin.
- A few feeds genuinely do not name an organiser. Those publish without one rather than us inventing a plausible name.
- A UK event outside our launch cities publishes without a city — findable by search, absent from city pages until regional pages ship.
How we treat the sites we read
We read public event pages and the feeds sites publish for the purpose — sitemaps, iCal, RSS, structured data. We honour robots.txt, we fetch slowly and with a cap, and where a site has said no, we do not have a listing from them.
We take the facts: what, when, where, how much. Descriptions are rewritten in our own words rather than copied. Every listing links back to the place you can actually book, and we take no cut of anything you buy there.
Where an organiser publishes their own event image we show it from their server, so their bandwidth and their credit stay theirs. We never copy it onto ours, never serve it from our domain, and never run it through a caching optimiser.
If you run a site or an event and you would rather not be here, say so and you are out. We honour that first and argue about it never.
Found something wrong?
Then one of the six gates let you down, and we would rather know. Every listing has a report link at the bottom; reports go into the same queue a person works through, not a black hole.