Eventbrite + Meetup Events Scraper

An Eventbrite scraper and a Meetup scraper in one Actor. Give it a city and a phrase and it returns event data from both platforms as one table: conferences, concerts, workshops, classes, networking nights and community meetups, each with a parsed venue address, coordinates, ticket prices, the organizer, and — where Meetup publishes it — the RSVP count.

No API key, no login, no proxy, no headless browser. Both platforms are read through the same public endpoints their own websites use, so a run costs a few requests and a few seconds rather than a browser session per page.

What this Actor is for

Event marketers building a calendar of everything happening in a city this month. Sponsorship and sales teams looking for conferences and trade nights worth showing up to. Community managers watching what competing meetups are running. Venue and ticketing analysts tracking ticket price ranges by category. Anyone who needs event listings as rows instead of as a web page.

The number that matters: one query stops at 1000 events

This is the single reason the Actor exists in this shape.

Ask Eventbrite for everything happening in New York and it answers, in its own pagination block, that it found 10000 events. Then it hands over 49 pages of 20. Page 50 still works. Page 51 returns an empty list — not an error, not a warning, just nothing. A scraper that trusts the object_count number and keeps paging quietly stops at a thousand rows and reports success.

Measured on 2026-09-12 against the live endpoint:

What we asked	What the source promised	What it actually hands over
All New York events, page 1	object_count 10000	page_count 49, 20 per page
Page 50	—	20 events
Page 51	—	0 events, no error

The way past it is that every date window gets its own cap. Walk the period one day at a time and each day is a fresh query with a fresh thousand. Same city, same endpoint, fourteen days:

Strategy	Events reachable
One plain query	1000
14 day windows	11802

That is 11.8 times more events out of the same source, with no proxy and no browser. The walk is on by default (breakPageCap); turn it off when you want the single fastest query and a thousand rows is enough.

The online catalogue has the same ceiling and the same cure: 50 pages, 1000 unique online events per query, walked day by day when you ask for a period.

Two platforms, one table

Every competing Actor in this niche scrapes one platform. This one scrapes two and merges them into a single schema, so platform is just a column. Rows from Eventbrite carry the ticket price range and the parsed postal address; rows from Meetup carry the attendance count and the size of the community behind the event. Fields a platform does not publish are null — never invented.

	Eventbrite	Meetup
Keyword search	yes	yes
Location	bounding box from your radius	latitude, longitude, radius
Date range	yes, and it drives the cap walk	yes
Category filter	20 categories, sent inside the request	not offered by the platform
Format filter	19 formats, sent inside the request	not offered by the platform
Free / paid filter	yes, sent inside the request	price returned, filtered here
Online events	separate catalogue (see limitations)	native filter
Attendance	not published	attendees_going
Community size	organizer followers, verified badge, events hosted	group_members
Full description	with includeDescription	always
Groups mode	—	yes

How the location works

You type Berlin, Austin, TX or Greater London. The Actor resolves it to coordinates first, then sends a bounding box to Eventbrite and the same distance in miles to Meetup. Ten cities on four continents were checked live and all ten resolved. An invented place resolves to nothing, and the run says so out loud instead of silently returning the whole planet — that failure has an empty_reason of its own. You can skip the lookup entirely by passing latitude and longitude.

Input

Field	What it does
searchQuery	Words to match. Empty means browse everything on Eventbrite.
location	City or place to search around.
latitude, longitude	Exact point, replaces the city lookup.
radiusKm	How far around the point to look, 1 to 500.
sources	eventbrite, meetup, or both. Empty means both.
mode	events, or groups for Meetup communities.
startDate, endDate	Period as YYYY-MM-DD.
daysAhead	Period length when no end date is given.
breakPageCap	Walk the period day by day to pass the 1000 cap.
categories	Twenty Eventbrite categories, sent inside the request.
formats	Nineteen Eventbrite formats: conference, concert, class and so on.
eventType	any, in_person, online, hybrid.
priceFilter	any, free, paid.
sortBy	Meetup order: relevance or date.
minAttendees	Drop events with fewer people going.
titleContains	Keep only events matching these words.
excludeKeywords	Drop events matching these words.
eventUrls	Scrape these exact Eventbrite event pages instead of searching.
includeDescription	Open each Eventbrite page for the full text and price range.
includeOrganizer	Add the organizer track record. One request per unique organizer.
maxItems	Stop after this many delivered rows. 0 means no limit.
onlyNew	Return only what is new or changed since earlier runs.
emitUnchanged	In that mode, also return unchanged events.
compactOutput	Seventeen columns instead of the full record.
excludeEmptyFields	Leave empty fields out of the JSON.

Output

Sixty-two columns, the same shape for both platforms.

Identity and text: platform, record_type, event_id, title, summary, description, description_html, url, tickets_url, image, tags.

Time: start_at, end_at, timezone, duration_minutes, published_at, scraped_at.

Place: venue_name, address, city, region, postal_code, country, neighborhood, latitude, longitude, is_online.

Money: is_free, price_min, price_max, currency.

People: organizer_name, organizer_url, organizer_id, organizer_followers, organizer_website, organizer_twitter, organizer_facebook, organizer_verified, organizer_total_events, organizer_upcoming_events, organizer_hosting_years, organizer_attendees_hosted, organizer_attendees_number, group_name, group_url, group_members, attendees_going, capacity.

Classification and state: category, subcategory, format, age_restriction, language, series_id, is_sold_out, status, is_cancelled, change_type, search_query, search_location, search_window.

Every column may be null, and the dataset schema says so. That is not sloppiness: Apify validates the dataset schema on a live run, and a column declared as a plain string that arrives empty fails the whole batch, so the buyer gets nothing.

The address is parsed, not a label

Eventbrite hands back the venue broken into street, city, region, postal code, country and coordinates, plus a place hierarchy that goes down to the neighbourhood. All of it is kept in separate columns. Meetup gives street, city, state, postal code and coordinates, and those land in the same columns. You can group by city, filter by postal_code or plot latitude/longitude without parsing a display string first.

Prices are numbers

price_min and price_max are numbers or null — never an empty string, never a marketing sentence. A column of floats with "" in it breaks Excel, BigQuery and pandas alike, so an unknown price is null and stays null. On Eventbrite the range comes from the ticket availability block, and includeDescription cross-checks it against the price schema published on the event page itself.

Who is behind the event

Turn on includeOrganizer and every Eventbrite row gains the organizer's track record: whether Eventbrite has verified them, how many events they have run in total, how many are still upcoming, how long they have been hosting, and how many attendees they have hosted. Together with the website, X and Facebook links that come free with the search, that is enough to sort organizers by seriousness before you contact any of them.

Two details worth stating plainly. The request is made once per unique organizer per run, not once per event, so a city sweep where one promoter runs forty parties costs one extra page, not forty. And the attendee count is published by Eventbrite already rounded, as 98k: the rounded string is kept in organizer_attendees_hosted exactly as shown, and organizer_attendees_number holds the same figure as an integer for sorting. Neither pretends to a precision the source does not have.

Monitoring without paying twice

Turn on onlyNew and the Actor remembers every event it has already delivered in a named key-value store, which survives between runs. The next run returns only what is new or what changed — a moved date, a renamed event, a cancellation. Unchanged events are counted and skipped, so a scheduled watch on a city costs a fraction of the first run.

The fingerprint behind that comparison is built only from fields that arrive in every mode: title, start time, venue and the cancellation flag. Description and price are deliberately excluded, because they only appear when includeDescription is on — including them would make the same event look "changed" every time you flipped that switch.

An incremental run that returns nothing is not a failure, and the Actor says which kind of nothing it is: the report carries an empty_reason of nothing-changed, filtered-out, unknown-location, source-returned-nothing or charge-limit-reached, and a matching warning goes into the log. Silence is the one thing it will not do.

maxItems limits delivery, not crawling

If maxItems stopped the crawl instead of the output, a filter could eat your entire order: the Actor would collect thirty rows, your filter would remove all thirty, and you would get an empty dataset from a green run. Here the crawl keeps going until the number of delivered rows is reached, with a bounded reserve so a rare filter cannot run up your bill either.

Duplicates are dropped before billing, not after. Meetup returns the same event on more than one page — measured: 157 rows arrived, 150 were distinct — and without deduplication you would pay for the difference.

Run report

Every run writes a REPORT record: how many day windows were walked, how many requests each platform took, how many raw rows arrived, how many were duplicates, how many each filter removed, whether a query hit the 1000 ceiling, and which platform failed if one did. When something returns less than you expected, the report says why.

Honest limitations

Things a competing Actor offers that this one does not, named rather than hidden:

• Eventbrite online events cannot be filtered inside the search request. The endpoint accepts online_events_only and silently ignores it — measured: twenty rows came back with the online flag false on every one. Choosing Online therefore switches Eventbrite to its online catalogue, which is a different listing with the same 1000-event ceiling. In that catalogue the source also refuses to combine a keyword with a date range: it drops the dates. The Actor puts the keyword in the request and applies your dates to the rows afterwards, so the filter you asked for is really applied.
• Meetup has no category taxonomy in event search. categories narrows Eventbrite only. Use searchQuery and titleContains for Meetup.
• Meetup sorts by relevance or date only. There is no sort by attendance; the platform does not offer one, and inventing it would mean sorting a page at a time and calling it global.
• Meetup returns nothing for an empty phrase. A neutral word is substituted and the log says so.
• Attendance is Meetup-only. Eventbrite does not publish how many people are going, so attendees_going is null there and minAttendees deliberately keeps those rows instead of deleting them.
• eventUrls accepts Eventbrite pages only. A Meetup event address is skipped and named in the log.
• formats and categories narrow Eventbrite only. Meetup has neither taxonomy in its event search.
• No country picker with 238 entries. Location is a place name or a point plus a radius, which is both more precise and easier to schedule.
• No built-in delivery to Notion or Slack. Use Apify's own integrations on the dataset.

Cost

Billing is per row delivered, so the filters above cost you nothing: rows removed before the dataset is written are never charged. The crawl itself is cheap because it is plain HTTP — one handshake page, then one request per page of twenty events, plus one extra request per event only if you asked for the full description.

FAQ

Do I need an Eventbrite or Meetup API key?

No. Both platforms are read through the public endpoints their own websites call. There is nothing to register and nothing to renew.

Does it need proxies or a browser?

No. The Actor is plain Python on the standard library, and it runs in 512 MB. That is why it is cheap enough to schedule hourly.

How do I get more than a thousand events?

Leave breakPageCap on and widen the period. Each day in the range is a separate query with its own thousand-event ceiling; fourteen days of New York reached 11802 events where a single query gives 1000. Narrowing by categories or by searchQuery also creates fresh caps, because the ceiling is per query, not per city.

Why does the report say hit_cap?

Because one of your queries found more events than the source will hand over. It is a hint, not an error: split the period into shorter windows, or add a category, and the same events become reachable.

Can I scrape one specific event page?

Yes. Put its address into eventUrls. Each page is read in full: description, price range, age restriction, duration, category, format and sales status.

Can I get Meetup groups instead of events?

Set mode to groups. You get community name, link, description, city, coordinates, member count, topic category and the founding date, in the same table shape with record_type set to group.

How often should I schedule it?

For a city watch, once or twice a day with onlyNew on. Event listings do not change by the minute, and incremental mode means you only pay for movement.

What happens if one platform is down?

The other one still runs. The failure is recorded in the report under that platform's name, and the log says what went wrong. A half-broken run returns half the data instead of nothing.

Is the data ordered?

Rows arrive in the order the platforms return them: Eventbrite first, window by window, then Meetup. Sort in the dataset view or downstream — start_at is a plain ISO timestamp.

Which categories are available?

Business & Professional, Science & Technology, Music, Film, Media & Entertainment, Performing & Visual Arts, Fashion & Beauty, Health & Wellness, Sports & Fitness, Travel & Outdoor, Food & Drink, Charity & Causes, Government & Politics, Community & Culture, Religion & Spirituality, Family & Education, Seasonal & Holiday, Home & Lifestyle, Auto, Boat & Air, Hobbies & Special Interest, and Other. All twenty were checked against the live source and all twenty return events.

Which formats are available?

Conference, Seminar or Talk, Tradeshow, Consumer Show, or Expo, Convention, Festival or Fair, Concert or Performance, Screening, Dinner or Gala, Class, Training, or Workshop, Meeting or Networking Event, Party or Social Gathering, Rally, Tournament, Game or Competition, Race or Endurance Event, Tour, Attraction, Camp, Trip, or Retreat, and Appearance or Signing. All nineteen formats were checked live and every one returns events.