Trade Show Exhibitor List Scraper | Company & Booth Data

Build CRM-ready exhibitor lists from supplied public trade show and conference directories. Each paid result is one qualified exhibitor record with company and booth data, source evidence, and optional public contacts when shown—not a guaranteed sales lead, navigation link, sponsor package, or generic page mention.

Current proof: a bounded IAEE directory probe on August 26, 2026 qualified 173 company/booth records, including 42Chat, 4imprint, and AEX Convention Services.

What you get

Each dataset row can include:

• event and company name
• booth or stand number
• company website, category, company description, email, phone, and LinkedIn only when present inside that exhibitor record
• source URL and extraction timestamp
• parser shape (table, card, or list), confidence, and the evidence signals used to qualify the row
• a stable deduplication key based on normalized exhibitor identity plus the source site's event URL scope (not the page title)

The buyer-facing paid unit is one qualified exhibitor record. The compatible internal event key remains lead-extracted; it is charged only after a record passes the qualification gate. maxItems is checked before every paid push, and processing stops immediately when the pay-per-event charge limit is reached.

Input

{
  "startUrls": [
    { "url": "https://s3.goeshow.com/iaee/annual/2026/exhbitor_list.cfm" },
    { "url": "https://www.aptaexpo.com/exhibitor_exhibitor_list.cfm" }
  ],
  "maxPages": 25,
  "maxItems": 100,
  "requestTimeoutSecs": 30
}

Field	Purpose	Limits
startUrls	Public exhibitor or vendor directory pages for a conference, expo, or trade show.	At least 1 URL
maxPages	Maximum supplied pages fetched.	1–500; default 25
maxItems	Maximum paid qualified leads emitted.	1–500; default 100
requestTimeoutSecs	Per-page timeout covering redirects and body consumption.	5–120; default 30

Only public HTTP(S) destinations are accepted. The fetcher checks DNS and every redirect, blocks loopback/private/link-local/reserved destinations, allows no more than five redirects, accepts HTML only, and rejects pages larger than 500,000 bytes rather than parsing truncated HTML.

Output

Example shape (illustrative field values; optional contacts appear only when the source record contains them):

{
  "eventName": "Expo! Expo! 2026",
  "companyName": "42Chat",
  "booth": "843",
  "sourceUrl": "https://s3.goeshow.com/iaee/annual/2026/exhbitor_list.cfm",
  "detectedAt": "2026-07-06T12:45:07.000Z",
  "dedupeKey": "42chat|s3.goeshow.com/iaee/annual/2026/exhbitor_list.cfm",
  "parserShape": "table",
  "confidence": "high",
  "evidenceSignals": ["exhibitor-directory-context", "booth-cell", "company-cell"]
}

Optional fields are website, companyDescription, category, email, phone, and linkedin. Missing source data remains absent; the Actor does not infer or fabricate contacts.

Supported page structures

The extractor qualifies these public directory shapes:

1. Booth-first tables with a booth/stand identifier in the first cell and company in the second, only after the table establishes exhibitor-directory context through semantic headers, a directory caption/class, or a nearby exhibitor heading.
2. Exhibitor or vendor cards with a named heading/link, explicit exhibitor/vendor directory context, and supporting booth or public-contact evidence.
3. List rows inside explicit exhibitor/vendor directory context, with a named heading/link and a booth/stand label.

Generic cards, schedules, navigation menus, organizer contact blocks, sponsor pricing/benefit/package sections, and standalone exhibitor keywords are not sufficient for a paid result. A sponsor-only heading, table, or CSS class is rejected even when its rows contain booth-like numbers. Client-rendered directories whose records are absent from the returned HTML are not covered by this lightweight HTML extractor. Login-only attendee lists, private data, documents, and non-HTML APIs are not supported.

Archived validation evidence (2026-07)

A July 2026 cloud validation used the public IAEE Expo! Expo! and APTA TRANSform & EXPO exhibitor pages at 256 MB. The combined run returned 778 rows. Dataset inspection confirmed booth-first examples including:

• 42Chat — booth 843
• 4imprint — booth 1335
• AEX Convention Services — booth 635

This evidence predates the current semantic-deduplication and sponsor-rejection hardening. It describes those archived page versions only; it is not current extraction evidence, a compatibility guarantee, or a claim that all 778 historical rows would qualify today. The current parser retains booth-first extraction only when the live table establishes exhibitor/vendor directory context.

Current bounded probe (2026-08-26)

A local probe from the built production image fetched one public IAEE URL with maxPages: 1 and a 30-second request timeout. The bounded fetch read 107,891 HTML characters and the hardened parser qualified 173 exhibitor rows. The first three were 42Chat (booth 843), 4imprint (booth 1335), and AEX Convention Services (booth 635), each with exhibitor-directory-context, booth-cell, and company-cell evidence. This is a point-in-time one-page probe, not a cloud run or a promise of future source availability.

Operating notes

• Use public pages you are authorized to access and follow the source site's terms and applicable law.
• Results reflect only the supplied URLs; the Actor does not discover every event directory on the web.
• A successful page fetch does not guarantee records if the markup lacks a supported, qualified shape.
• Large or script-rendered directories may require a browser-based extractor instead.
• Keep sourceUrl, evidenceSignals, and dedupeKey in downstream workflows for review and deduplication.
• Semantically duplicate exhibitors are suppressed across pagination, page-title, and extracted eventName variants in the same run before charging. Event-distinguishing URL path/query components remain in the key so separate events on the same host do not collapse.

Development

The Actor runs on Node.js 24 with npm 12.0.2. Install the committed dependency graph and run the local checks with:

npm ci
npm run lint
npm test
npm run typecheck

Use npm start to run the TypeScript entry point through the pinned tsx runtime.