Official scripts
Explore our curated collection of scripts
Descript whoami (signed-in account)
Report which Descript account this browser is signed in as: the user id, email and name. Being signed out is reported as a normal answer (loggedIn false), not an error, so it can be used to check a session before running other Descript scripts.
Descript login (Google SSO)
Log into Descript via 'Continue with Google'. Drives the chooser and confirms the logged-in web.descript.com app shell. Only works when run through the browser extension, using the browser's native Google session — it does not work on managed/cloud browser runs.
Descript whoami (signed-in account)
Report which Descript account this browser is signed in as: the user id, email and name. Being signed out is reported as a normal answer (loggedIn false), not an error, so it can be used to check a session before running other Descript scripts.
List Descript invoices
List Descript billing invoices for the signed-in drive. Returns id, number, date, total (in cents), currency, paid and receiptUrl (a Stripe receipt page linking to the invoice PDF) per invoice.
Get Dribbble shot
Fetch a Dribbble shot's details from its url: title, description, owner, images and comment count.
Search Dribbble shots
Search Dribbble shots by keyword, returning title, url, owner, thumbnail, likes and views for each result.
Get Booking.com hotel details
Fetch a hotel's Booking.com listing: name, star rating, guest rating, review count, address, description, and image. No login required.
Maps: Search Businesses
Search for businesses on Google Maps (anonymous, no login): name, category, address, phone, website, rating, reviews, GPS, place_id. Paginates via scrolling, controlled by the count argument.
Search G2 software products
Search G2's software directory by free-text query. Returns matching products with slug, name, star rating, review count and description, deduplicated. The slug feeds g2.com/get_product. Whitespace in the query is normalised. No login required. Fails loudly rather than returning an empty list: G2 always degrades to fuzzy matches, so zero results means the page did not render.
Search Google Maps places
Search Google Maps for places matching a query, optionally near a location. Returns n, captured, and records with name, address, phone, website, rating, category, lat, lng, placeId, and placeUrl. placeId is the join key for get_place and get_place_reviews; rating, category, lat and lng can be null.
Search Wellfound jobs
Search Wellfound (AngelList Talent) startup job listings by role keyword, with optional city location (omit for remote-only). Returns total, page, totalPages, and jobs (jobId, title, url, company, companyUrl, jobType, compensation, location, experience, postedAgo). Unrecognized roles or locations raise an error rather than returning unrelated results. The reported total can exceed the number of job rows actually returned on a page.
Get Similarweb competitors and similar sites
Read a website's competitors / similar sites from its public Similarweb overview page: competitor domain, affinity percent, category and category rank. No login required. Returns the ~10 competitors the anonymous page renders; the section's "See all competitors" expander leads to gated Similarweb product pages, so that is the anonymous ceiling. available:false means Similarweb has no report for that domain.
Like YouTube video
Like a YouTube video — has no effect if it's already liked.
Search Pinterest boards
Searches Pinterest boards by keyword and returns the first page of results (name, url, pin count, cover images, owner). Read-only, no login required; Pinterest serves a reduced anonymous result set.
Search Pinterest pins
Searches Pinterest pins by keyword and returns the first page of results (id, pin URL, thumbnail/full image, dominant color). Read-only, no login required; Pinterest serves a reduced anonymous result set with no title/description per card. Be aware that a query with no real matches still returns Pinterest's recommended pins rather than an empty list.
Get Malt freelancer profile
Fetches a Malt freelancer's full public profile by seoAlias (the slug from search_freelancers' profileUrl, e.g. "samuelguetta"). Returns bio, address, availability, price, languages, category, experiences, educations, certifications and visible client reviews. Languages/categories are sometimes omitted by Malt's own page even for real profiles (page-state variance), so treat empty as a legitimate result, not an error.
Dealroom: Get Company Profile
Fetch a company's full Dealroom profile by its Dealroom path/slug (as returned by the Search Companies script) — overview, HQ, funding rounds, investors, valuation, industries, technologies and social links.
Download brand logo pack
Given a company domain, download its full Brandfetch logo pack (logos, icons and banners as SVG/PNG/WebP) as a zip. Returns the brand name and the file; found is false when Brandfetch has no page for the domain.
Claude Console login (Google SSO)
Log into the Claude Console (platform.claude.com, formerly console.anthropic.com) with Continue with Google. Returns straight away if you are already signed in, and reports the account.
Get Pinterest pin details
Fetches a single Pinterest pin's public details by id: title, description, image, creator, save/reaction counts, publish date, and (for link pins) the linked destination site. Read-only, no login required.
Search Booking.com hotels
Search Booking.com for accommodations by destination: name, hotel slug, address, guest rating, review count, and rating category. No login required.
Search Claude chats
Search your Claude.ai chats (the "Search chats…" box): server-side full-text over titles + message bodies. Each hit carries the matched snippet and match offsets the UI bolds.
Search Google
Run a Google search and return the page's organic result cards (title, url, site, byline, meta, snippet), with redirect wrappers already cleaned from urls. Runs logged out by default; if the browser happens to carry Google cookies, results are personalized to that account/IP. Google serves result links as redirects rather than destination addresses, so each one is followed far enough to read the address it points at before the results are returned. That adds a fraction of a second per result, and each destination site sees one request. Supports full Google query syntax: site:, quoted "exact phrases", OR, -exclude, intitle:. Quoted phrases return a smaller, more precise set; dropping quotes broadens recall with some near-miss results mixed in. Use site: to scope a domain/path, e.g. site:www.linkedin.com/in. Each call returns one results page (up to ~10 organic cards, no per-call count option). Use `page` (0-based) to move through further pages, mirroring the search results pager. Pages are reproducible on the same account/IP and are largely non-overlapping (occasional overlap of one result at page boundaries), so different pages can be fetched concurrently and unioned — always dedupe by canonical url. Narrow queries return the same results in the same order across replays on one account/IP; very broad, high-volume queries may reorder between runs. For broader coverage, fetch multiple pages concurrently and/or split the query itself — vary the site: path, phrases/OR-terms, or use `freshness` for disjoint date-bounded windows. Always dedupe by url.
Get Trustpilot company reviews
Get one page (20) of a company's Trustpilot reviews: author, rating, title, text, dates, likes, verification and the company's reply. Supports the site's own filters as args (page, sort, stars, language, keyword search, verified, date range). Defaults to returning reviews in every language so none are silently filtered out by locale.
Search talent.com jobs
Search talent.com's aggregated job listings by free-text keyword and optional free-text location (city/region; omit for a broad/geo-default search). Returns richer job data (salary, company, remote flag, job type, description) than what's shown on the page, read directly from the page's own underlying data rather than scraped visually — more stable against markup changes. Paginates via the site's own page param (page=1 default, ~18-20 results/page); no total-count field is exposed anywhere, so callers must page until an empty array comes back. Location text is matched loosely against talent.com's own place index (e.g. "paris" can resolve to Paris, TX, US rather than Paris, France) — this mirrors the site's own search box behavior, not an added restriction. A query with no matches returns jobs:[] — not an error.