BROWSER USE - Browser Use Agents: give Browser Use a task and receive completed work. API V4 is current for new integrations. - Browser Infrastructure: connect your agent or automation to managed browsers through SDK, REST, or CDP. Starts at $0.02/browser-hour. - Developer tools: Open Source, Browser Harness, SDK, and MCP support the two products above. [Developer Index](https://browser-use.com/index.md) [Product Map](https://browser-use.com/llms.txt) [Full Product Context](https://browser-use.com/llms-full.txt) [Pricing](https://browser-use.com/pricing.md) [Cloud Docs](https://docs.browser-use.com/cloud/quickstart) [Open Source Docs](https://docs.browser-use.com/open-source/introduction) --- # Ticketmaster scraper skill for AI agents Source: https://browser-use.com/skills/ticketmaster. One file from github.com/browser-use/browser-harness, MIT licensed. Written for browser-harness: code calls its helpers (http_get, js, goto_url). Sites change; these notes describe what worked when last updated. --- # Ticketmaster - Events by keyword or category (name, date, venue, event URL) Field-tested against www.ticketmaster.com on 2026-11-01 using a Browser Use cloud browser (US exit IP), logged out. No CAPTCHA, queue or block page appeared in about 40 page loads at 2 s or more spacing. robots.txt (read in the browser on 2026-11-01, `User-agent: *`) has no rule matching `/search?q=`, `/discover/...` or `//event/`, and disallows `/*event/*?preselect`, `/*event/*&preselect`, `/search*searchVariant=`, `/search*f_discoAiSearch=`, `/json/search/event/*`, `/api/ismds/event`, `*did=` and the checkout, member and queue paths; this skill only opens `/search?q=...&sort=...`, `/discover/...` and plain event URLs, and checks every URL against the file before loading it. ## Quick summary - **Every list page ships its first 20 events in `__NEXT_DATA__`.** `/search?q=` has `searchEvents(...)`, `/discover/` has `categoryEvents(...)`, `/discover/` has `cityEvents(...)`, all under `props.pageProps.initialReduxState.api.queries`. Each event has `title`, `id`, `url`, `dates.startDate` (UTC), `timeZone` and `venue` (`name`, `city`, `state`, address, lat/long). - **More pages = the page's own "More Events" button.** Clicking it makes the page fetch `/api/search/events?...&page=N` (search), `/api/search/events/category/?page=N` (category) or `/api/search/events/city?page=N...` (city), 20 events each, same shape. The block clicks the button and reads that response through CDP `Network` events; it never requests those URLs itself. - **Filters are plain URL parameters** written by the page's own controls: `sort=relevance|date` on search; `classificationId=` and `startDate=YYYY-MM-DD&endDate=YYYY-MM-DD` on `/discover/...` (the "This Weekend" button wrote exactly that). - **Event pages** come in two builds (`/globalevent/[eventId]` and `/naevent/[eventId]` in `__NEXT_DATA__.page`); both ship one schema.org Event JSON-LD block with name, local start time, venue and address, so the block reads that. - **Many list rows are partner events** (10 of 20 on `/discover/chicago`: TicketWeb, hotel packages) whose `url` points off-site; `on_ticketmaster` is `False` for them and `event_details()` refuses them. ## What works: search, category, city and event pages in one run Self-contained; opens one tab (robots.txt first, then every page in the same tab) and closes it at the end. About 1-3 minutes in a cloud browser (48 s on 2026-11-01). ```python import json, re, time from datetime import datetime from urllib.parse import urlencode, urlparse from zoneinfo import ZoneInfo BASE = "https://www.ticketmaster.com" # robots.txt, group "User-agent: *", Google-style matching: longest pattern wins, Allow wins a tie. # Read in the browser tab: plain urllib got HTTP 403 for robots.txt (curl got 200), and this skill does not change its UA. _TAB = [] def _robots_rules(): new_tab(BASE + "/robots.txt") _TAB.append(current_tab()) wait_for_load() txt = js("document.body.innerText") if "User-agent" not in (txt or ""): raise RuntimeError("could not read robots.txt; stop") rules, in_star, prev_ua = [], False, False for line in txt.splitlines(): line = line.split("#", 1)[0].strip() if ":" not in line: continue k, v = (x.strip() for x in line.split(":", 1)) k = k.lower() if k == "user-agent": in_star = (in_star and prev_ua) or v == "*" prev_ua = True continue prev_ua = False if in_star and k in ("allow", "disallow") and v: rules.append((k == "allow", v)) return rules RULES = _robots_rules() def robots_allowed(path): best_len, ok = -1, True for allow, pat in RULES: rx = re.escape(pat.rstrip("$")).replace(r"\*", ".*") + ("$" if pat.endswith("$") else "") if re.match(rx, path) and (len(pat) > best_len or (len(pat) == best_len and allow)): best_len, ok = len(pat), allow return ok def _path(url): u = urlparse(url) return u.path + ("?" + u.query if u.query else "") def open_page(path): """Open an allowed www.ticketmaster.com page and return its __NEXT_DATA__ (the state the page ships in its HTML).""" if not robots_allowed(path): raise PermissionError(f"robots.txt disallows {path}") goto_url(BASE + path) # the tab opened for robots.txt deadline = time.time() + 30 while time.time() < deadline: try: # event pages can keep the main thread busy for >5 s, so a single evaluate may time out raw = js("(() => { const s = document.getElementById('__NEXT_DATA__');" " return document.readyState !== 'loading' && s ? s.textContent : null })()") href = js("location.href") except RuntimeError: raw = None if raw: final = _path(href.split("#")[0]) if not robots_allowed(final): raise PermissionError(f"{path} redirected to disallowed {final}") return json.loads(raw) time.sleep(1) info = page_info() raise RuntimeError(f"no page state on {info['url']} ({info['title']!r}); a queue, block or challenge page? " "Stop here, do not retry around it.") def _row(e): d, v = e.get("dates") or {}, e.get("venue") or {} start, tz = d.get("startDate"), e.get("timeZone") local = None if start and tz and not e.get("tba"): local = datetime.fromisoformat(start.replace("Z", "+00:00")).astimezone(ZoneInfo(tz)) url = e.get("url") or "" return { "id": e.get("id"), "name": e.get("title"), "local_date": local.strftime("%Y-%m-%d") if local else None, "local_time": local.strftime("%H:%M") if local and d.get("dateDisplay") == "showDateTime" else None, "date_display": d.get("dateDisplay"), # showDateTime has a time; other values show the date only "start_utc": start, "time_zone": tz, "venue": v.get("name"), "city": v.get("city"), "state": v.get("state"), "country": v.get("countryCode"), "url": url, # Partner events (TicketWeb, hotel packages, ...) link off-site; only www.ticketmaster.com/.../event/ is an event page here. "on_ticketmaster": urlparse(url).netloc == "www.ticketmaster.com" and "/event/" in url, "status": ("cancelled" if e.get("cancelled") else "postponed" if e.get("postponed") else "rescheduled" if e.get("rescheduled") else "sold_out" if e.get("soldOut") else "listed"), } LIST_KEYS = ("searchEvents(", "categoryEvents(", "cityEvents(") def _first_page(nd): q = nd["props"]["pageProps"]["initialReduxState"]["api"]["queries"] hits = [(k, v["data"]) for k, v in q.items() if k.startswith(LIST_KEYS) and v.get("data")] if len(hits) != 1: raise RuntimeError(f"expected one event list in the page state, found {[k for k, _ in hits]}") return hits[0] def _retry(fn, tries=6): """Ad-heavy pages can keep the tab busy for a few seconds; a CDP call then times out. Retry a few times.""" for i in range(tries): try: return fn() except (RuntimeError, TimeoutError): if i == tries - 1: raise time.sleep(2) def _more_events_button(): for n in _retry(lambda: cdp("Accessibility.getFullAXTree")["nodes"]): if (n.get("role") or {}).get("value") == "button" and (n.get("name") or {}).get("value") == "More Events": return n["backendDOMNodeId"] return None def _click_more_and_read(timeout=20): """Click the page's own "More Events" button and return the JSON the page itself fetches for it.""" node = None for _ in range(8): # the button renders after hydration node = _more_events_button() if node: break time.sleep(1) if node is None: return None _retry(lambda: cdp("DOM.scrollIntoViewIfNeeded", backendNodeId=node)) q = _retry(lambda: cdp("DOM.getBoxModel", backendNodeId=node)["model"]["content"]) drain_events() click_at_xy(sum(q[0::2]) / 4, sum(q[1::2]) / 4) rid, deadline = None, time.time() + timeout while time.time() < deadline: for ev in drain_events(): p = ev.get("params") or {} if ev.get("method") == "Network.responseReceived" and p["response"]["url"].startswith(BASE + "/api/search/events"): if p["response"]["status"] != 200 or not robots_allowed(_path(p["response"]["url"])): raise RuntimeError(f"More Events got {p['response']['status']} for {p['response']['url']}; stop here") rid = p["requestId"] elif rid and ev.get("method") == "Network.loadingFinished" and p.get("requestId") == rid: return json.loads(_retry(lambda: cdp("Network.getResponseBody", requestId=rid)["body"])) time.sleep(0.2) raise RuntimeError("More Events was clicked but no event list response arrived") def list_events(path, max_pages=1, pause_s=2.0): """Events from a /search?q=... or /discover/... page: page 1 from the page state, later pages through the page's own "More Events" button (20 per page). Returns (rows, total, status).""" nd = open_page(path) _retry(lambda: cdp("Network.enable")) key, data = _first_page(nd) total, rows, seen = data.get("total"), [], set() for page in range(1, max_pages + 1): new = [r for r in map(_row, data.get("events") or []) if r["id"] not in seen] seen.update(r["id"] for r in new) rows.extend(new) if len(seen) >= (total or 0): return rows, total, "complete" if page == max_pages: return rows, total, "truncated_max_pages" time.sleep(pause_s) data = _click_more_and_read() if data is None: return rows, total, "no_more_button" def search(keyword, sort="relevance", max_pages=1): """Keyword search. sort: relevance | date (the page's own Sort by options).""" return list_events("/search?" + urlencode({"q": keyword, "sort": sort}), max_pages) def discover(category, classification_id=None, start_date=None, end_date=None, max_pages=1): """Category pages: concerts | sports | arts-theater | family, or a city slug such as new-york-city. classification_id = a sub-genre id from the category menu; dates as YYYY-MM-DD (what the page's date filter writes).""" params = {k: v for k, v in (("classificationId", classification_id), ("startDate", start_date), ("endDate", end_date)) if v} return list_events(f"/discover/{category}" + ("?" + urlencode(params) if params else ""), max_pages) def event_details(url): """Name, date, venue and status from an event page (/.../event/). Refuses non-Ticketmaster URLs.""" u = urlparse(url) if u.netloc != "www.ticketmaster.com" or "/event/" not in u.path or u.query: raise ValueError(f"not a plain www.ticketmaster.com event URL: {url}") nd = open_page(u.path) # Two page builds exist (__NEXT_DATA__.page "/globalevent/[eventId]" with eventInfo, "/naevent/[eventId]" with # edpData); both ship one schema.org Event JSON-LD block, so read that. lds = json.loads(js("JSON.stringify([...document.querySelectorAll('script[type=\"application/ld+json\"]')]" ".map(s => { try { return JSON.parse(s.textContent) } catch (e) { return null } }))")) ev = next((x for x in lds if isinstance(x, dict) and x.get("startDate") and x.get("location")), None) if not ev: raise RuntimeError(f"no Event JSON-LD on {page_info()['url']}") loc, addr = ev.get("location") or {}, (ev.get("location") or {}).get("address") or {} perf = ev.get("performer") or [] return { "id": u.path.rstrip("/").rsplit("/", 1)[-1], "name": ev.get("name"), "start_local": ev.get("startDate"), # venue-local wall time, no offset: "2026-11-01T18:15" "status": (ev.get("eventStatus") or "").rsplit("/", 1)[-1] or None, # EventScheduled, EventCancelled, ... "venue": loc.get("name"), "address": ", ".join(str(x).strip() for x in (addr.get("streetAddress"), addr.get("addressLocality"), addr.get("addressRegion"), addr.get("postalCode")) if x), "performers": [p.get("name") for p in (perf if isinstance(perf, list) else [perf]) if isinstance(p, dict)], "availability": ((ev.get("offers") or {}).get("availability") or "").rsplit("/", 1)[-1] or None, "page_build": nd.get("page"), "url": page_info()["url"], } def show(r): when = f"{r['local_date']} {r['local_time'] or ''}".strip() return f"{when} | {r['name'][:60]} | {r['venue']}, {r['city']} {r['state'] or ''} | {r['url']}" # 1. Keyword search, sorted by date, 2 pages (page 2 via the page's own "More Events" button) rows, total, status = search("jazz", sort="date", max_pages=2) print("search jazz:", len(rows), "of", total, status) for r in rows[:3] + rows[20:22]: print(" ", show(r)) time.sleep(2) # 2. Category page with a sub-genre and the date filter (Country concerts, next Saturday-Sunday) from datetime import date, timedelta sat = date.today() + timedelta(days=(5 - date.today().weekday()) % 7 or 7) rows2, total2, status2 = discover("concerts", classification_id="KnvZfZ7vAv6", start_date=sat.isoformat(), end_date=(sat + timedelta(days=1)).isoformat(), max_pages=2) print(f"discover concerts/Country {sat}..{sat + timedelta(days=1)}:", len(rows2), "of", total2, status2) for r in rows2[:3]: print(" ", show(r)) time.sleep(2) # 3. City page rows3, total3, status3 = discover("chicago") print("discover chicago:", len(rows3), "of", total3, status3, "| off-site:", sum(not r["on_ticketmaster"] for r in rows3)) time.sleep(2) # Check: most on-site event URLs end in the venue-local date (MM-DD-YYYY); it must equal local_date. # The rest are slug-less /event/ URLs with no date in them. from collections import Counter dated = [r for r in rows + rows2 + rows3 if r["on_ticketmaster"] and r["local_date"] and re.search(r"-\d\d-\d\d-\d{4}/event/", r["url"])] slug_ok = sum(f"-{r['local_date'][5:7]}-{r['local_date'][8:10]}-{r['local_date'][:4]}/event/" in r["url"] for r in dated) print("slug date == local_date:", slug_ok, "of", len(dated), "| date_display:", Counter(r["date_display"] for r in rows + rows2 + rows3)) time.sleep(2) # 4. Event pages: the first two on-site search results, plus one slug-less /event/ URL if listed picks = [r for r in rows if r["on_ticketmaster"]][:2] picks += [r for r in rows2 + rows3 if r["on_ticketmaster"] and urlparse(r["url"]).path.startswith("/event/")][:1] for r in picks: d = event_details(r["url"]) print(json.dumps(d, ensure_ascii=False)) print(" list row said:", r["local_date"], r["local_time"], "| page says:", d["start_local"]) time.sleep(2) # 5. A query with no results print("nonsense query:", search("zzqxqvnotathing")[1:]) # 6. Disallowed paths are refused before the tab navigates for p in ("/search?q=jazz&searchVariant=ai", "/json/search/event/foo", "/some-show/event/0B00653E2B2EC4EE?preselect=1"): print("allowed?", p, robots_allowed(p)) close_tab(_TAB[0]) # the one tab this block opened ``` Output observed on 2026-11-01 (trimmed): ```text search jazz: 40 of 3889 truncated_max_pages 2026-10-31 21:00 | Halloween with Sally Baby's Silver Dollars | Snug Harbor Jazz Bistro, New Orleans LA | https://www.ticketweb.com/event/halloween-with-sally-babys-silver-snug-harbor-jazz-bistro-tickets/14410584 2026-11-01 15:00 | Fresh Takes On Jazz: Ava Preston | Mickey's Black Box, Lititz PA | https://www.ticketmaster.com/fresh-takes-on-jazz-ava-preston-lititz-pennsylvania-11-01-2026/event/0200651B00735077 2026-11-01 18:00 | Johnny Britt "You Are Beautiful" Tour | Humphreys Backstage Live, San Diego CA | https://www.ticketmaster.com/johnny-britt-you-are-beautiful-tour-san-diego-california-11-01-2026/event/0A0065159E4494E2 discover concerts/Country 2026-11-07..2026-11-08: 40 of 108 truncated_max_pages 2026-11-06 20:00 | Saddle Up: Country Grammar | Diamondback Music Hall, Belleville MI | https://www.ticketweb.com/event/saddle-up-country-grammar-diamondback-music-hall-tickets/14407284 2026-11-07 12:30 | BACKSTAGE NASHVILLE! DAYTIME HIT SONGWRITERS SHOW featuring | 3rd and Lindsley, Nashville TN | https://www.ticketweb.com/event/backstage-nashville-daytime-hit-songwriters-3rd-and-lindsley-tickets/14637713 discover chicago: 20 of 2425 truncated_max_pages | off-site: 10 slug date == local_date: 32 of 32 | date_display: Counter({'showDateTime': 100}) {"id": "0200651B00735077", "name": "Fresh Takes On Jazz: Ava Preston", "start_local": "2026-11-01T15:00", "status": "EventScheduled", "venue": "Mickey's Black Box", "address": "101 Rock Lititz Blvd, Lititz, PA, 17543", "performers": [], "availability": null, "page_build": "/naevent/[eventId]", "url": "https://www.ticketmaster.com/fresh-takes-on-jazz-ava-preston-lititz-pennsylvania-11-01-2026/event/0200651B00735077"} list row said: 2026-11-01 15:00 | page says: 2026-11-01T15:00 {"id": "010064CB19BF6123", "name": "Huntertones", "start_local": "2026-11-01T19:30:00", "status": "EventScheduled", "venue": "Jimmy's Jazz and Blues Club", "address": "135 Congress Street, Portsmouth, NH, 03801", "performers": ["Huntertones"], "availability": "InStock", "page_build": "/globalevent/[eventId]", "url": "https://www.ticketmaster.com/huntertones-portsmouth-new-hampshire-11-01-2026/event/010064CB19BF6123"} list row said: 2026-11-01 19:30 | page says: 2026-11-01T19:30:00 {"id": "Z7r9jZ1AAZd8O", "name": "Jon Pardi w/ Alexandra Kay", "start_local": "2026-11-07T17:30", "status": "EventScheduled", "venue": "Round Rock Amp", "address": "3701 N IH 35, Round Rock, TX, 78665", "performers": [], "availability": null, "page_build": "/naevent/[eventId]", "url": "https://www.ticketmaster.com/event/Z7r9jZ1AAZd8O"} list row said: 2026-11-07 17:30 | page says: 2026-11-07T17:30 nonsense query: (0, 'complete') allowed? /search?q=jazz&searchVariant=ai False allowed? /json/search/event/foo False allowed? /some-show/event/0B00653E2B2EC4EE?preselect=1 False ``` ## URL patterns (all observed) | URL | Page state key | Notes | |---|---|---| | `/search?q=` | `searchEvents({"keyword":..,"page":0,"sort":"relevance"})` | `total` = all matches (310 for `taylor`, 3889 for `jazz`) | | `/search?q=&sort=date` | `searchEvents({...,"sort":"date"})` | soonest first; includes events that started earlier today or yesterday | | `/discover/concerts`, `/discover/sports`, `/discover/arts-theater`, `/discover/family` | `categoryEvents({"categoryId":..,"page":0})` | the four top menu categories (`headerConfig(undefined).data.categories`) | | `/discover/?classificationId=` | `categoryEvents({"categoryId":"",...})` | sub-genre; ids are in `categoryInfoV2(...).data.minor` and the menu's `subCategories` (Country = `KnvZfZ7vAv6`) | | `...&startDate=YYYY-MM-DD&endDate=YYYY-MM-DD` | `categoryEvents({...,"startDate","endDate"})` | the page's date filter; works on direct load | | `/discover/` (`new-york-city`, `chicago`, ...) | `cityEvents({"cities":..,"stateCodes":..})` | 23 city slugs in `headerConfig(undefined).data.cities` | | `/-/event/<16 hex>` | `eventInfo` or `edpData` | Ticketmaster-hosted event page | | `/event/` (e.g. `Z7r9jZ1AAZd8O`) | `edpData` | slug-less event page, same JSON-LD | ## Fields | Output field | Source | Notes | |---|---|---| | `id` | `events[i].id` | 16 hex chars, a discovery id such as `Z7r9jZ1AAZd8O`, or a partner id (`14720223`) | | `name` | `title` | | | `local_date`, `local_time` | `dates.startDate` converted with `timeZone` | `startDate` is UTC (`...Z`) or has an offset (`-05:00`); every on-site URL with a date in its slug matched (32/32) | | `date_display` | `dates.dateDisplay` | all 100 rows sampled were `showDateTime` | | `venue`, `city`, `state`, `country` | `venue.name/city/state/countryCode` | `venue` also has `addressLineOne`, `code` (ZIP), `latitude`, `longitude`, `url` | | `url`, `on_ticketmaster` | `url` | off-site rows link to ticketweb.com or travel.ticketmaster.com | | `status` | `cancelled`, `postponed`, `rescheduled`, `soldOut` | `listed` when none is set | | event page `start_local` | JSON-LD `startDate` | venue-local, no offset; the two builds format it differently (`T15:00` vs `T19:30:00`) | | event page `status`, `availability` | JSON-LD `eventStatus`, `offers.availability` | `offers` was missing on the `/naevent/` pages tried | ## Gotchas - Category and city lists are sorted by start date and begin with long-running listings that started months ago (season tickets, museum passes, hotel packages; on 2026-11-01 the first `/discover/concerts` row was a tribute show dated 2025-11-21). Filter on `local_date` if you want upcoming events only. - `total` was exactly 10000 for `/discover/concerts`, `/discover/sports` and `/discover/new-york-city`; it looks like a display cap, so read 10000 as "10000 or more". - The date filter is not strict local dates: `startDate=2026-11-07` returned a 2026-11-06 20:00 event in Michigan (01:00 UTC on 11-07) and multi-day listings that started earlier. Re-check `local_date` yourself. - Search rows carry a `local` flag, but it is not a distance signal: on 2026-11-01 all 20 first-page rows for `taylor`, `jazz` and `comedy` had `local: true`, with venues in NJ, DC, MI, CA, UT, AZ and more. - The "PROMOTED" card shown between search results is an ad slot and is not in `searchEvents`; the code never sees it. - Ad-heavy pages sometimes keep the tab busy for over 5 s, so single CDP calls (`DOM.getBoxModel`, `Runtime.evaluate`) time out; the block retries them. Close other heavy tabs in the same browser. - If a page has no `__NEXT_DATA__` (a queue, block or challenge page), the code raises. Stop there; do not retry, switch IP or change the request to get around it. ## What does not work - Plain `urllib` with its default User-Agent got HTTP 403 for `/robots.txt` (curl got 200), so the block reads robots.txt in the browser tab instead of changing its User-Agent. - `/api/ismds/event` (seat and price availability) and `/json/search/event/*` are robots-disallowed; this skill does not request them. Event pages load the former themselves; the block does not read it. - Event URLs with `?preselect=` and search URLs with `searchVariant=` or `f_discoAiSearch=` (the AI search) are robots-disallowed and refused by `robots_allowed()`. - Partner rows (TicketWeb, travel.ticketmaster.com) are other sites with their own robots.txt; `event_details()` refuses them.