Eventbrite Event Scraper

Scrape upcoming Eventbrite events by city and category — title, dates, venue, price range and organizer name — with no personal data, robots.txt respected and rate-limited crawling.

What it does

Give it a city (and optionally a category), get back upcoming Eventbrite events as clean, flat rows: title, start/end date, timezone, venue name, city, country, online flag, price range, category, organizer name and URL, the event URL, and the browse page the event was found on (sourceListUrl).

Discovery works through Eventbrite's own sitemap and category browse pages (/b/<city>/<category>/), which are stable and reachable from the Apify platform. City names are resolved against the sitemap, so "San Francisco", "Istanbul" or the exact slug ca--san-francisco all work. If a name is ambiguous (springfield, or new york as city vs. state) the run stops and lists the candidates instead of guessing.

{
  "city": "San Francisco",
  "category": "music",
  "maxItems": 20,
  "futureOnly": true
}

Output

One row per event. Every field is present in every row; missing values are null, never omitted.

{
  "id": "1983682844788",
  "url": "https://www.eventbrite.com/e/prime-the-san-francisco-meat-2026-tickets-1983682844788",
  "title": "PRIME - The San Francisco MEAT 2026",
  "description": "PRIME - The San Francisco MEAT!",
  "startDate": "2026-09-26T05:00:00.000Z",
  "endDate": "2026-09-26T11:00:00.000Z",
  "timezone": "-07:00",
  "venueName": "Club Six - Both Floors!",
  "city": "San Francisco",
  "country": "US",
  "isOnline": false,
  "price": { "min": 75.38, "max": 75.38, "currency": "USD" },
  "category": "San Francisco Music Events",
  "organizerName": "TrophyDad",
  "organizerUrl": "https://www.eventbrite.com/o/trophydad-11423585232",
  "attendeeCount": null,
  "sourceListUrl": "https://www.eventbrite.com/b/ca--san-francisco/music/",
  "scrapedAt": "2026-09-14T19:05:25.010Z"
}

Notes on fields:

• attendeeCount is always null today: Eventbrite does not expose it in the page's structured data.
• timezone is the UTC offset of the event's local time (-07:00), not an IANA name.
• description is truncated to 2,000 characters; emails, phone numbers and URLs inside it are masked.
• sourceListUrl tells you which browse page the event was found on — useful for auditing and for understanding why an event from another city can appear (Eventbrite occasionally lists it there).

Input

Request budget for the example input above: 32 requests (1 sitemap index + 5 sitemap shards + 6 browse pages + 20 event pages). The run log prints the planned and actual request counts.

Category slugs

Categories are Eventbrite's own browse slugs and differ slightly per city. If you pass one that does not exist for the city, the run stops and lists the valid ones. For San Francisco (measured 14 Sep 2026) they are:

arts, auto-boat-and-air, business, charity-and-causes, community, family-and-education, fashion, film-and-media, food-and-drink, government, health, hobbies, holiday, home-and-lifestyle, music, nightlife, other, school-activities, science-and-tech, seasonal, spirituality, sports-and-fitness, tech, travel-and-outdoor

What it does NOT do

• No organizer discovery. Organizer pages (/o/…) render their event lists client-side; the static HTML carries no events, so this Actor does not crawl them. organizerName and organizerUrl come from the event page itself.
• No personal data. Attendee names, emails, phone numbers, social handles and profile pictures are never collected. Free-text fields are scanned and masked before export. Organizer name is taken only when Eventbrite marks it as an organization, not a person.
• Date filters are post-filters. dateFrom / dateTo narrow the output, not the crawl: they do not reduce the number of requests and do not widen the search.
• No /d/ list pages. Eventbrite's /d/ search listings answer datacenter traffic with a human-verification challenge about three times out of four, and the outcome changes between attempts. This Actor does not retry those pages. If you pass a /d/ URL in startUrls it is tried once and skipped.
• No ticket purchase, no login, no API endpoints. Only public pages that Eventbrite's robots.txt allows.

Limits and compliance

• Personal data: none collected. Output is a fixed whitelist of 18 fields; anything else on the page is dropped. Free text is masked for emails, phones and URLs.
• robots.txt: fetched once per run and obeyed. Paths it disallows are never requested. If robots.txt cannot be fetched after three attempts, the run stops rather than guessing.
• Rate limiting: 2 concurrent requests, at most 60 requests per minute, 0.5–1.5 s random delay before every request. Human-verification responses are never retried; five in a row stop the run.
• Coverage: browse pages list at most 8 events each. A single city + category typically exposes a few hundred upcoming events; ask for more than exist and you get a warning with the exact numbers.
• Freshness: data is read live at run time. Eventbrite's sitemap does not carry per-event modification dates, so there is no "changed since" mode.
• Failures are loud. If discovery cannot start, if a record fails schema validation, or if the internal counters do not reconcile, the run fails with a diagnostic message. It never finishes green with an empty dataset and no explanation.

FAQ

Why not use the Eventbrite API?

The public search endpoint was retired in 2020. What remains requires an API key, OAuth, and in practice a partnership for anything beyond your own events. This Actor reads public pages instead — no credentials, no application process.

Why did an event from another city appear in my results?

Browse pages are Eventbrite's own listings, and they occasionally surface events from nearby or unrelated places. The city field comes from the event's own structured data, not from your input. Use sourceListUrl to see which browse page produced the row.

Can I get organizer emails or attendee lists?

No, by design. This Actor collects no personal data. If you need contact details, this is not the right tool.

What happens if I ask for more events than exist?

You get the ones that exist, plus a warning stating how many were requested, how many were found, how many browse pages were scanned, and why there are no more.

Pricing

Pay per result: $3.50 per 1,000 results ($0.0035 per event row). You pay only for rows written to the dataset. A run that finds nothing costs only platform usage (roughly $0.007 for a 20-event run at 1 GB memory on the reference runs).