Data collected June 2026 across 17 cities. Updated continuously.

TL;DR


The number that surprised us

When we built hiddenevents.online, the goal was simple: show all Luma events, not just the featured ones. We assumed the gap would be small — maybe Luma missed 20-30% of events.

We were wrong.

When we started crawling organizer calendars across San Francisco, Bengaluru, Singapore, and other cities, the pattern was consistent: the discover page was showing a small fraction of what actually existed.

Not a rounding error. Not a technical glitch. A structural gap, built into how Luma works.

Here’s the data.


City-by-city: how many events Luma buries

We pulled this data from our live crawler, which checks Luma’s public APIs every 4 hours across 17 cities:

City Featured Hidden Total % Hidden
San Francisco 82 1202 1284 94%
Singapore 31 125 156 80%
Bengaluru 19 136 155 88%
New York City 65 236 301 78%
Boston 11 52 63 83%
Mumbai 9 19 28 68%

In SF, you’re seeing roughly 1 in 6 events on the discover page. The other five? You’d have to follow every organizer individually to know they exist.

The pattern holds across cities with very different event volumes. It’s not a San Francisco anomaly — it’s how Luma’s discover algorithm works everywhere.


What “hidden” actually means

A clarification, because this is counterintuitive:

Hidden events aren’t private. They’re fully public. Anyone with the link can view them, RSVP, and attend. There’s no paywall, no approval gate. If you somehow got the URL, you’d be at the event.

What’s hidden is discoverability. These events don’t appear on lu.ma/discover. They don’t show in search results on Luma. They exist only in the organizer’s calendar page — which you’d have to know to look for.

The mechanics: Luma gives every organizer a calendar page (e.g., lu.ma/yourorgnamehere). Every event they post appears there. But Luma’s discover page is a curated selection — not a complete feed. The curation criteria aren’t fully public, but from our data, the pattern is clear: size and organizer reputation dominate.

A new community calendar with 20 followers posting a genuine, high-quality founder dinner? Hidden. A corporate sponsor with 5,000 followers posting a paid conference? Featured.


Why the hidden ones are usually better

This is what makes the gap matter.

The events that don’t get promoted are almost always the events worth attending:

Founder dinners — Usually 10-20 people, curated by someone with actual judgment about who’s building interesting things. These are almost never featured. The host picks guests personally; they don’t need Luma’s algorithm to fill the room.

Niche AI reading groups — Paper-reading clubs, LLM fine-tuning circles, agentic systems discussions. Small, recurring, no marketing budget. The people who show up are genuinely working on the thing, not watching it from a distance.

Recurring weekly meetups — The ones that have been running every Tuesday for two years. The Luma discover page surfaces new and trending. Recurring events are invisible by design.

Community calendars — A local Rust community, a product design collective, a biotech founders group. These organizers are running events for their community, not for growth. They don’t optimize for Luma’s algorithm. Their events disappear into the long tail.

The 5,000-person AI conference with a keynote and a live stream gets featured. The 30-person demo night where three founders are showing actual working agents doesn’t.


Why Luma works this way

Luma isn’t trying to hide events. The discover page is a product decision, not a malicious filter.

Luma’s business model depends on event organizers using the platform. Featured placement is a growth lever — it rewards organizers who are driving engagement and RSVPs. Featured events build the organizer’s reputation on Luma, which drives more follows, which drives more RSVP velocity, which keeps them on Luma.

The discover page is optimized for platform retention, not for attendee discovery. These aren’t the same thing.

For attendees, comprehensive discovery would mean seeing every event in your city. For Luma’s growth, comprehensive discovery means showing you the events most likely to make you RSVP, come back, and subscribe to more organizers.

The result is a curation that skews toward whatever gets RSVPs fastest — which is usually scale (big existing audiences) and paid (promoted listings), not quality.


What you’re actually missing

To make this concrete: in San Francisco right now, there are 1202 events on Luma that you can’t find from the discover page.

Some of those are duds — private meetups that got accidentally set to public, canceled events that weren’t removed, niche gatherings that wouldn’t interest you. But many aren’t.

In the last 30 days of data from our crawler across all cities:

The events you find on the Luma discover page are the ones Luma wants you to see. The events worth finding are often somewhere else on the same platform.


How to actually find them

Three approaches, in order of effort:

1. Use hiddenevents.online — We crawl Luma every 4 hours across 17 cities and surface everything: featured and hidden, sorted by date. Filter by city, category (AI, startup, tech, networking), or time (today, this weekend). Each event links straight to Luma for RSVP. Free, no account needed.

2. Follow organizer calendars directly — If you’ve been to a good event, follow the organizer on Luma (lu.ma/discover → their profile → Follow). Their future events will appear in your “following” feed, not just discover. Build a list of 10-15 organizers worth following and you’ll see the good hidden events before they fill up.

3. Find the city’s meta-community — Most active tech cities have a Slack, WhatsApp group, or Discord where event organizers share upcoming events. SF has several. Bengaluru’s tech WhatsApp groups are extremely active. These communities surface hidden events that even our crawler might take a few hours to catch. Ask in the first event you attend — someone will know where the signal is.


The data setup

For those interested in methodology:

We use a crawler that pulls from Luma’s public graph APIs — the same APIs that power lu.ma’s own web app. We collect every event from every organizer calendar we’ve indexed, not just what appears on the discover page. We classify an event as “hidden” if it doesn’t appear in Luma’s featured/discover endpoint but does appear in the organizer’s calendar endpoint.

Our city index currently covers 17 cities across North America, Europe, India, and Southeast Asia. We update every 4 hours. The crawler has been running since early 2026, giving us roughly six months of longitudinal data — enough to see the pattern isn’t seasonal.

The 80% figure is a weighted average across cities. Individual city rates currently range from 68% (Mumbai) to 94% (SF, which has the highest event volume and the most competitive featured placement).


What this means for organizers

If you run events on Luma, the hidden rate is useful signal:

Getting featured is not the only goal. A solid organizer following of 200-500 engaged people will fill your events reliably without needing featured placement. Focus on building your calendar subscribers, not gaming Luma’s discover algorithm.

Your hidden events still get found. People who follow your calendar see everything you post. People who use aggregators like ours see it within hours. The discover page is one channel — not the only one.

Recurring events get buried. Luma’s discover page weights toward newness and trending RSVPs. If you run a recurring weekly meetup, it will rarely get featured no matter how good it is. Compensate with direct follow growth and cross-platform promotion.


If you’re looking for events in your city — browse the full live list →