Extract AD PRO Interior Designer Leads

Purpose

Extract contact "leads" for interior designers listed in the AD PRO Directory on architecturaldigest.com. For each business this returns: firm name, contact person name (when named), email, phone, and website. It is a read-only scraping task — no login, no forms submitted, no state changed. Every field lives in the initial server-rendered HTML, so the recommended method is plain HTTP fetch (no browser, no proxies, no anti-bot handling required).

When to Use

You need a list of interior-design firms from the AD PRO Directory with their public contact details (email, phone, website).
You want to build a lead/outreach list scoped to a directory category (e.g. Interior Design & Decor, or by extension Architecture, Builders & Contractors, Outdoor & Garden Design).
You need firm-level contact info for a specific AD PRO profile slug you already have.
You do not need this if you want editorial article content — this skill targets the business directory only.

Workflow

Recommended method: HTTP fetch of server-rendered HTML. The listing and profile pages return complete HTML on a plain GET (HTTP 200, no cookies, no JS execution, no proxy needed). There is no JSON/GraphQL API to call — all data is in the HTML.

Fetch the category listing, page 1. GET https://www.architecturaldigest.com/adpro/directory/businesses?category=interior-design-decor Read the results header to get the total count, e.g. 1 - 50 of 1,241 Results. Results are 50 per page, so total_pages = ceil(total / 50) (≈25 pages for interior design).
Paginate. Append &page=N (1-based) to walk every page: .../businesses?category=interior-design-decor&page=2, &page=3, … up to total_pages.
Collect profile links from each page. Each business card links to a profile path matching /adpro/directory/profile/{slug}. Extract the unique slugs (expect ~50 per page). The firm name also appears next to each card, but the authoritative fields come from the profile page in the next step.
Fetch each profile page and parse the header contact block. GET https://www.architecturaldigest.com/adpro/directory/profile/{slug} and extract:
- firm_name — the page <title>, formatted "{Firm} | Architectural Digest" (strip the | Architectural Digest suffix).
- website — the href of the anchor with class/attr HeaderAddressWebUrl (an external URL). May be absent → null.
- email — the mailto: target of the anchor with class/attr HeaderAddressEmail.
- phone — the tel: target of the anchor with class/attr HeaderAddressPhone. Stored literally as +(1) 312 644 9270 (with parentheses and spaces) — normalize downstream if you need E.164.
- person_name — not a structured field. Derive best-effort from the About/description prose ("principal designer X", "founder X", "led by X") and/or the email local-part (steve@…, liz@…). Set null when no individual is named (many firms use info@…).
- Optional extras also present: HeaderAddressInstagram, and address parts HeaderAddressStreet/City/State/PostalCode/Country.
Assemble one lead object per profile and emit the array (see Expected Output). Throttle politely — a full category sweep is ~25 listing fetches + ~1,241 profile fetches.

Browser fallback

If you must drive a browser (e.g. the fetch path is unavailable in your runtime), it works but is slower/costlier:

Open the category URL in a remote session. Wait for load.
Do not use browse snapshot — it intermittently errors (Error: error taking snapshot: -3) on these heavy Verso pages. Use browse get markdown body or browse get html body for both the listing (to harvest /adpro/directory/profile/{slug} links) and each profile (to read the contact block).
A cookie-consent overlay (Fides, id="fides-overlay") prepends "Manage your consent preferences…" text to the extracted markdown — ignore that preamble; the real data (phone/email/website/firm) is below it. You do not need to dismiss the overlay to read the DOM.
Reuse the single open session — repeatedly issuing browse open --remote triggers Session "default" is already running and wastes turns.

Site-Specific Gotchas

No data API — do not look for one. The directory and profile data are fully server-side rendered. The only XHR/fetch traffic on these pages is third-party analytics/ad pixels (reddit, twitter/x, bing, doubleclick, pinterest). Parse the HTML.
No anti-bot / no auth. Homepage probe reported no antibots. Both proxied and bare (no proxy, no --verified) fetches returned HTTP 200 with full content. Residential proxies are not required. No AD PRO login is needed to read directory profiles.
Category slug is interior-design-decor. The directory mixes disciplines; without the filter you get all 1,616 businesses. Other observed category slugs: architecture, builders-and-contractors, outdoor-garden-design. Filter via ?category=interior-design-decor.
Pagination is &page=N, 50 results/page. It combines with the category filter: ?category=interior-design-decor&page=2. Interior Design & Decor had 1,241 results (~25 pages) as of Aug 2026; counts drift over time, so always read the live N of M Results header rather than hardcoding.
Profile URL shape. Cards link to /adpro/directory/profile/{slug}. A bare /profile/{slug} path also appears in metadata but the canonical, fetchable page is under /adpro/directory/profile/{slug}.
person_name is unreliable/optional. It is not a first-class field — it only appears in prose. Firms frequently list a generic info@ mailbox and name no individual; return null rather than guessing.
Phone formatting is quirky. The tel: value is stored as +(1) 864 250 0021 (literal parentheses). Keep the raw value and normalize separately if needed.
<title> suffix. Firm name comes as "{Firm} | Architectural Digest" — strip the suffix. There is also a generic <title id="…">Architectural Digest</title> node elsewhere in the doc; take the first, descriptive <title>.
Browser snapshot is flaky here. browse snapshot returned -3 errors on these pages; prefer text/markdown/HTML extraction. (Documented in the browser fallback above.)
Scale/politeness. A full interior-design sweep is ~1,266 requests. Add delays/concurrency limits; the data is stable enough to cache profile HTML.

Expected Output

A JSON object with pagination metadata and an array of lead objects. website and person_name are nullable.

{
  "success": true,
  "source": "AD PRO Directory",
  "category": "interior-design-decor",
  "total_results": 1241,
  "results_per_page": 50,
  "pages_fetched": 1,
  "leads": [
    {
      "firm_name": "Kadlec Architecture + Design",
      "person_name": "Steve Kadlec",
      "email": "steve@kadlecdesign.com",
      "phone": "+(1) 312 644 9270",
      "website": "https://www.kadlecdesign.com",
      "profile_url": "https://www.architecturaldigest.com/adpro/directory/profile/kadlec-architecture-design"
    },
    {
      "firm_name": "Liz Caan & Co.",
      "person_name": "Liz Caan",
      "email": "liz@lizcaan.com",
      "phone": "+(1) 617 244 0424",
      "website": "http://www.lizcaan.com",
      "profile_url": "https://www.architecturaldigest.com/adpro/directory/profile/liz-caan-co"
    },
    {
      "firm_name": "Panageries",
      "person_name": null,
      "email": "info@panageries.com",
      "phone": "+(1) 864 250 0021",
      "website": "https://panageries.com/",
      "profile_url": "https://www.architecturaldigest.com/adpro/directory/profile/panageries"
    }
  ],
  "error_reasoning": null
}

Per-lead schema:

{
  "firm_name": "string        // from <title>, ' | Architectural Digest' stripped",
  "person_name": "string|null // best-effort from prose/email local-part; null if none named",
  "email": "string|null       // mailto: from HeaderAddressEmail",
  "phone": "string|null       // tel: from HeaderAddressPhone; literal e.g. '+(1) 312 644 9270'",
  "website": "string|null     // href from HeaderAddressWebUrl; null if not listed",
  "profile_url": "string      // canonical /adpro/directory/profile/{slug} URL"
}

On failure (e.g. an unexpected block or a category with zero results), return success: false with leads: [] and a populated error_reasoning.

extract-designer-leads