SnoopScan

Check whether a page has changed since it was last fetched.

Everything a company's own website says about itself. Give it a company domain and get back the firmographics a lead list sells — name, phone, address, LinkedIn, headcount, industry, founded year — plus the contact emails, social links and contact form the site publishes. Reads the homepage and a few contact/about pages, nothing behind a login. Use it to enrich one company. It is the single-company half of a lead pipeline, not a directory search.

Start crawling a website. This runs in the background and returns a job id immediately. Use `crawlStatus` to check progress and `crawlPages` to read results. Crawls can take minutes and consume significant resources — set `limit` conservatively.

Read the pages a crawl has collected, a few at a time. Small defaults on purpose — a crawl's full output will not fit in context.

Check the progress of a crawl. Returns counts and cost, never page bodies.

How old a domain is, who registered it, where it is hosted, and who links to it. The off-page questions a page scrape cannot answer. Use it when a page looks fine and you need to know whether the SITE is old, established or linked to — comparing competitors, judging a source, or checking whether a domain was registered last week. Nothing here fetches the page, so nothing here can be blocked.

A buyer's report on a domain: key findings, concerns, and the questions nobody can answer from outside, each tied to its evidence. `depth` is `quick` (registration, DNS, today's site), `buyer` (adds the Internet Archive history, known backlinks and a re-read sample of linking pages) or `deep` (larger samples). Brand, SEO and acquisition are judged separately; there is no overall score. One flat price per depth; `buyer` includes 6 archived snapshots and `deep` 12.

Which of up to 500 domain names can be registered, from each registry's own record (RDAP) and DNS. Use it to cut a long list of candidate or dropping names down to a shortlist, then run domainReport on the few that matter. Nothing is fetched from the names' websites. Each name comes back `likely_available`, `registered`, `unknown` or `no_registry_record`, with the reason. 1 credit per name its registry answered for; invalid names, duplicates and unanswered names are free.

Read the archived homepage snapshots a `buyer` domainReport left pending. A buyer report answers at once with the Internet Archive's timeline and the snapshots already held; the rest are marked pending. Call this with the same domain (and optionally the pending years) to read them: what the site was about in each year, its language, and whether it was parked, for sale or redirecting. Free for the snapshots a buyer or deep report included; beyond those, charged only for snapshots actually fetched.

Extract structured data from one or more pages according to a JSON schema. The output is validated against the schema — if a page does not contain the required fields, that page returns an error rather than invented values. Treat an error as genuine absence, not a reason to retry.

Read the remainder of a page that was truncated. Pass the token from a truncation marker. Returns the next section of the same page, truncating again if it is still too long.

Find contact information for a company website. Returns the contact page, any contact form and its vendor, published email addresses with where each was found, and social links. Used by the lead pipeline.

Find businesses by what they do and where, with how to reach each one. `who` is the trade or kind of business ("roofing contractors", "dentists"), `where` is a town or area ("Houston, TX"). Looks on Google Maps, BBB and Yellow Pages by default (`sources` may also name companies_house for UK companies by trade, or web); merges duplicates; reads each business's own website for emails, socials and a contact form (`contacts`); and with `roles` (e.g. ["Owner"]) looks for a named person. Runs in the background for 2 to 5 minutes: call `leadsStatus` with the job id to follow it and to read the leads. Charged per lead delivered, never for what it cannot find.

What a company is currently recruiting for. A company hiring six engineers has budget, a roadmap and a problem — the signal sales teams and recruiters pay most for. Give it a company domain. It finds the careers page, identifies the applicant tracking system, and reads that system's own public job feed, so it returns real roles rather than the empty shell most careers pages serve to a plain fetch. Not every company exposes a feed. When none is found it says so and names the careers page instead of pretending — do not read "no feed" as "not hiring".

Follow a Find Leads run, and read its leads once it has finished. Returns the stage while it runs. When done, returns the leads 25 at a time: name, phone, email, website, address and where each was found; call again with `offset` for the next 25.

List a site's posts or articles from its own API or feed. WordPress, Substack, Squarespace and Discourse answer through their APIs; anything else through RSS or Atom. Titles, URLs and dates — use `scrape` on a URL to read one.

List every product a store publishes, from the store's own catalogue. Works for Shopify and WooCommerce stores without a key: one request per page of up to 250 products, with title, price, availability, variants and the product URL. Use this instead of crawling a shop.

List the URLs on a website without fetching page content. Fast and cheap. Use this to understand a site's structure before deciding what to scrape.

Find the people at a company, by what they do and how senior they are. Head-hunting a developer, poaching a sales team, or finding who signs the cheque are the same shape of question: function plus seniority. Give it a company and optionally a `role` and a `seniority` and it returns named people from public search — no login, no profile scraping. role: executive, engineering, product, sales, marketing, finance, procurement, talent, operations, data, security, legal, support, design seniority: exec, vp, director, manager, ic Each person says whether the employer was CONFIRMED in the source text. Unconfirmed people are real leads but unproven — the company name did not appear beside them — and anyone marked former has LEFT. Check those two flags before you contact anybody.

Fetch a single web page and return its content as clean markdown. Use this when you have a specific URL and need to read what is on it. For finding pages first, use `search` or `map`.

Search the web and return results. Optionally fetch the content of each result. Fetching content is much slower and more expensive — only set `fetchContent` when you actually need the page bodies rather than just the links and snippets.