Browser WebMCP tools for AI agents

Extract contact leads for interior designers listed in the Architectural Digest AD PRO Directory — firm name, contact person (when named), email, phone, and website — by paginating the Interior Design & Decor category and parsing each server-rendered profile page.

Crawl docs.stagehand.dev via its Mintlify llms-full.txt export, resolve the linked GitHub source repo (browserbase/stagehand) pinned to a commit SHA, and use Cerebras (Qwen3-Coder-480B) to flag drift in function signatures, flag/option names, types, and example code. Emits a JSON report with docs URL + source file/line citations. Read-only.

Search 1688.com (Alibaba's China-domestic wholesale platform) by keyword and return structured offers — title, CNY wholesale price, MOQ, supplier name + location, recent-transaction count, and canonical offer URL. Supports price-range, supplier-province, and sort filters. Headed browsers are hard-blocked at IP level (cloud_ip_bl); the recommended path is the mtop JSON API at h5api.m.1688.com with md5-signed requests.

Return current real-time MPH for every California highway covered by ABC7's traffic map — per road, per direction, with active incidents — by querying the Sigalert/Total Traffic Network JSON backend that the abc7news.com/traffic/ page embeds via iframe.

Given a date, return every Nippon Professional Baseball (NPB) game on that day with final score, both teams + team IDs, venue, winning / losing / save pitchers, game status, and the canonical Yahoo game_id — parsed from baseball.yahoo.co.jp's weekly schedule page.

Collect every team member from blockchaintelligence.com (BIPA) — name, role, profile URL, photo, and full bio — via the WordPress REST API roster page plus per-member profile-page fetch. Read-only.

Find Sea Ranch vacation-rental homes on booksearanchescape.escapia.com by date range, party size, bedrooms, and pet allowance. Returns availability (day-by-day JSON), rate quotes, amenities, and detail-page links. Read-only — never books.

List Browserbase's starter templates (TypeScript / Python / Go) and the canonical clone command for each, sourced from the official bb CLI's templates subcommand which reads github.com/browserbase/templates@dev.

Fetch an AEO (Answer Engine Optimization) ranking / AI-visibility report for a user-supplied website from crawlproof.com. As of 2026-05-19 the target domain is a parked Railway deployment returning 404 on all paths, so the skill currently emits a service_unavailable result and surfaces aeoproof.com as the closest live alternative.

Given a Cruise Critic ship page (URL, or cruise line + ship name resolved via search), extract ship metadata plus a filtered slice of member reviews as structured JSON, leading with a Browserbase stealth session and using the per-review Next.js data JSON endpoint as a per-review enrichment shortcut.

Return Derek Meegan's most recently published articles (titles, dates, tags, and canonical Medium URLs) from the /writing index on derekmeegan.com. Read-only; single HTTPS GET, no browser or anti-bot stealth required.

Enumerate the ASID Design Finder Designers & Firms directory and extract each listing's name, contact person, email, phone (when published), and website.

Given a free-text query about a Scout platform feature, concept, integration, or API/SDK reference, locate the relevant page on docs.scoutos.com and return structured JSON with title, breadcrumb, headings, prose, code blocks, tables, On-This-Page anchors, last-updated, canonical URL, and related pages.

Extract a Malaysian real-estate agent's display name, full mobile phone (E.164), and email from any EdgeProp.my agent profile or listing page by parsing the inlined Next.js __NEXT_DATA__ payload — bypasses the UI's masked phone and email-form gateway.

Extract a company's overall rating, sub-rating averages, recommend/outlook/CEO-approval percentages, and a filtered slice of employee reviews from Glassdoor — accepting a URL, EmployerId, or company name (+ optional location disambiguator) and supporting the full review-page filter rail (stars, job title, location, employment status, language, employment type, sort, keyword, limit). Read-only.

Log into HiringCafe and open the user's Saved Jobs page (the HiringCafe Application Tracker at /myhiringcafe/tracker). Documents the email/password and Google/Microsoft/Facebook/GitHub OAuth login surface and the account-gated tracker.

Retrieve details for University of Huddersfield academic programs — list every course or fetch one course's level, duration, start dates, delivery mode, entry requirements, fees, and modules — via the coursefinder's SearchStax/Solr JSON API, with a prerendered-page and browser fallback.

Search ImmobilienScout24 for apartments to rent in the Köln (Cologne) district of Rodenkirchen, filtered by living space, room count, and monthly rent, returning the matching listings. Browser-driven; the result pages sit behind an AWS WAF wall that blocks datacenter/proxy automation.

Return JustWatch's daily 'Nuevo' feed for Spain (/es/nuevo) grouped by streaming platform, with each title's IMDb score, vote count, TMDB score, and Rotten Tomatoes meter pulled from the page's embedded Apollo cache — no per-title page visit needed.

Look up Liveblocks developer documentation for any topic and return the page title, URL, and full content. Uses the site's machine-readable docs (llms.txt index + per-page .md endpoints) instead of scraping rendered HTML.

Summarize Magic Eraser's AI photo-editing tools (object removal, background removal, enhance, expand, fill, filters, text-to-image, design) and its availability on web, iOS, and Android.

Fetch Hacker News stories from any list view (front, newest, ask, show, jobs, best, active, classic, by-domain, by-user, historical day) and optionally full comment trees, returning a unified JSON shape.

Teaches an agent how to navigate Notion (entry points, the login wall, workspace anatomy, navigation primitives) and what functionality each kind of page exposes — doc pages, blocks, and databases with their view types.

Search and compare OpenRouter's 350+ LLMs by cost, speed (throughput/latency/uptime), context length, modalities, and use-case category; pick the best fit; then build with it via OpenAI-compatible chat completions. API-first — no scraping, no auth for reads.

Extract iFLYTEK's open source project catalog (name, homepage, category, description, languages, tags, GitHub stars/forks, license, repo URL) plus aggregate ecosystem stats from opensource.iflytek.com/projects.

Extract structured property information from a realestate.com.au listing or address — location data (address + lat/lon), price range or sold price, key statistics (beds/baths/parking/land size/property type), agent + agency, and the full historical sale price timeline from the Property Pages address profile.

Extract StreetEasy for-sale listings (NYC + NJ) matching multi-dimensional filters — price, beds/baths, sqft, year built, property type, amenities, and area — returning structured per-listing data via the JSON-LD path. Bypasses PerimeterX by using headless HTTP fetch instead of a full browser session.

Play a live round of The Wiki Game: plan a short Wikipedia link-path from the round's Start article to its Goal with the Wikipedia API, then click through it in the browser within the 120s clock, returning the winning path, clicks, and score.

Search TheWindowsClub for articles by keyword (with optional category/tag scoping and date/relevance sort), returning each article's title, URL, publish date, excerpt, category and tag IDs.

Given a US street address, route to the assigned polling place + drop-box / early-voting alternatives via Vote.org's state-directory page-data.json (recommended) or the polling-place-locator HTML directory (fallback). Read-only; never registers or starts a check-in flow.

Discover Wander vacation-rental listings (via sitemap/locations) and extract a property's complete metadata plus the full image/video gallery from its detail page using server-rendered JSON-LD and the RSC medias arrays — read-only, no auth, no anti-bot.

Look up an .au domain on the auDA WHOIS service (whois.auda.org.au) to read registrant/business name, registrar name, and eligibility (ABN/ACN) details. Read-only. Access is gated by a two-stage reCAPTCHA Enterprise that reliably steps automated sessions up to an unsolvable image challenge.

Authenticate a user session against Xero — either via OAuth 2.0 / OIDC (recommended, supported) or as a fallback by scripting the password form at login.xero.com/identity/user/login. Documents the canonical URL, form schema, anti-bot stack (Akamai + browsercheck + AspNetCore antiforgery), and all five branch outcomes (MFA, SSO, passkey, lockout, invalid credentials).

Briefing of the four above-the-fold modules on the Yahoo! JAPAN homepage: 8 main news topics, 5 realtime trending search keywords with stay/rise/fall direction, today/tomorrow weather for the IP-defaulted ward, and the day's NPB scoreboard. Single static HTML fetch — no JS, no auth, no anti-bot.