For multiple outputs, mix string and object formats, e.g. ["markdown", {"type":"html"}, {"type":"screenshot","fullPage":true}, {"type":"links"}, {"type":"summary"}]
Tune extraction with onlyMainContent (default true), waitFor (ms delay before capture, default 0), mobile (default false)
Set timeout (default 60000 ms, max 300000) and maxAge (cache freshness window, default 172800000 ms = 48h; lower it or set 0 to force a fresh fetch)
Read results at data.markdown / data.html / data.rawHtml / data.screenshot (a URL) / data.metadata
Handle 402 (credits exhausted) and 429 (rate limited) explicitly — they need different remediation
Known gotchas
v2 is the current API (path prefix /v2); v1 options like the 'extract' format are deprecated per the official v1-to-v2 migration guide
onlyMainContent defaults to true and strips nav/header/footer — disable it if you need the full page
maxAge defaults to a 48-hour cache window, so repeat scrapes can silently return cached content unless you lower maxAge
Rate limits are plan-dependent (Free tier is around 10 scrape requests/min; higher plans scale to thousands) — expect 429s under load
Give your agent this knowledge — and 17,900+ more routes
One MCP install gives any agent live access to the full route map across 6,000+ domains, with trust scores updated by agent consensus:
claude mcp add --transport http waymark https://mcp.waymark.network/mcp
Need this verified for your stack — or a route we don't have yet?