No rentedproxy pool.Nothing to rotate.
askFinz doesn't buy a residential proxy pool and doesn't rotate through addresses it doesn't control. Its fleet reads through its own VPN exits in several countries — its own infrastructure, identifiable, and accountable when something goes wrong.
reaching a site/
├── rented proxy pool/
│ ├── whose addresses someone else's
│ ├── identity anonymous
│ └── accountable to nobody
└── askfinz/
├── whose addresses our own fleet
├── exits several countries, ours
├── identity askFinz-Crawler/2.0
└── accountable to crawler docs, a humanOur own exits. Our own name. Our own problem when it goes wrong.
Our own fleet, its own exits
Reading runs on infrastructure askFinz operates — the same accountable fleet described on how the index works — and its connections route out through VPN exits in several countries rather than one fixed address. There is no third-party residential proxy pool being rented or rotated behind it.
A 403 means the IP was refused, not that the site is unreachable
A refusal is information, not a dead end. A blocked request is re-queued carrying the exits already refused, and it's only marked failed once every distinct egress available to it has been turned away — never given up on after one no.
Dead ends heal instead of piling up
A page that stays unreachable is tracked in a pool-wide dead-URL registry rather than being retried forever — and it isn't written off permanently either. It's reconsidered after 14 days, so a page blocked today isn't blocked in the record for good.
Identifiable, on purpose
Requests carry askFinz-Crawler/2.0 and a link back to a documentation page a site owner can actually read. A rented, rotated proxy pool exists to be anonymous. That isn't what this is for — see the crawler documentation for exactly how to identify, allow or block it.
A 403 says one exit was refused. It doesn't say the site is unreachable.
A rented residential proxy pool is built to make traffic look like it's coming from ordinary people's home connections, at scale, from an address you don't control and usually can't see. That is a different thing from a fleet with its own exits, and conflating the two would be describing infrastructure askFinz doesn't operate.
A refused IP is more useful as a fact than a problem to launder around. Knowing that a specific exit was turned away is what lets a retry pick a genuinely different one and eventually tell the difference between "this exit was refused" and "this page doesn't exist."
A page only ends up marked unreachable once every distinct exit available to the fleet has been turned away — and even then it isn't final. It sits in a dead-URL registry that reconsiders it after 14 days, rather than treating one bad day as a permanent verdict.
Does askFinz use proxies at all?
It reads through its own fleet's VPN exits in several countries — not through a rented, rotating residential proxy pool bought from a third party. That's a meaningful difference: the exits are ours, accountable, and identifiable, not an anonymous pool sold by the connection.
What actually happens when a site returns a 403?
It's treated as "this particular exit was refused," not "this site can't be read." The request is re-queued along with the list of exits already turned away, and only counted as a real failure once every exit available has been refused.
Do blocked pages get retried forever?
No — a page that stays unreachable moves into a pool-wide dead-URL registry so it isn't hammered on repeat, and that registry self-heals: a page is reconsidered after 14 days rather than being blocked in the record permanently.
How can a site owner identify or block this traffic specifically?
Every request carries the User-Agent askFinz-Crawler/2.0 with a link to the crawler documentation, which covers exactly how to allow or block it — match on that token, not on an IP address, since the exit address is expected to vary.
You decide how it gets read.
Allow it, slow it down, feed it in directly as a partner, or ask us not to read the site at all — the crawler documentation covers all four, and none of them requires an account.