FAQ

Is this “random strangers crawl for me”?

Not in mesh v1. Mesh v1 is deliberately designed so you run a private mesh of known operators (--network-id + --allowed-peer) and can require approval for risky jobs.

Do I get JavaScript rendering?

Only if an operator supplies a browser adapter. The worker does not bundle a browser runtime by default. That’s intentional for safety and packaging. In v1 the adapter is a trust boundary: treat it like “a separate service that can click and run JS”.

What’s the main value for v1 users?

• Ship features faster: write queries, not bespoke scrapers.
• Stop re-scraping: cache-first results + explicit @fresh to refetch.
• Safer defaults: caps/backoff + deny POST/submit/cross-origin unless explicitly enabled.
• Split responsibilities: “API gateway” vs “execution workers” (and optional validators).

Can it crawl a whole site?

v1 supports a bounded crawl across discovered links (<a href>) with hard caps and per-origin backoff. It’s meant for “crawl up to N pages and stop when a target field appears”, not for unlimited whole-web crawling.

Does it support XPath and pagination?

Yes on the basics. v1 extraction supports CSS selectors and an XPath‑lite subset. For pagination, you typically extract links (like “next page”) and either follow them via a bounded crawl or a small plan (in browser mode if the site needs JS).

Does it do logins, CAPTCHAs, or “bypass bot protection”?

No. v1 is intentionally conservative: it’s a safer fetch/extract tool, not an anti-bot circumvention kit. If you need logins, you’ll likely need a custom adapter/operator setup and you should consider the legal/ethical risks.

Is it safe to open-source?

Open source doesn’t eliminate misuse risk. The practical mitigation is to ship with strong defaults, clear operator controls, and a mesh posture that can be tightened.