Government portals, property registries, court records, business directories. Rotating datacenter IPs through one endpoint, anti-detect TLS, and a rendered browser for JavaScript-only search forms.
Paste the portal you actually need to scrape. See the real response before you give us anything.
Public records are public. They sit behind rate limits and session state, not behind consumer-grade bot detection. Paying $4–8/GB for residential exits to read a county assessor's database is money set on fire — you are buying stealth the target never asked for.
BotProxy gives those targets what they actually need: volume, geographic coverage, and a pool that has not been burned by fraud traffic. At datacenter economics.
500 GB/month costs $300 here. The same volume on mid-market residential runs $1,000–2,000.
Six workloads account for most of the traffic here. Each one has its own page on what blocks it and how to configure for it.
Scraping public data at scale fails in predictable ways, and most of them cost money quietly rather than breaking loudly. The answer to the biggest one is the price: datacenter economics for targets that were never protected against you in the first place. The rest of this list is what we built after watching the other failures happen.
We manage datacenter IPs in every location we sell, and size each pool to the load actually on it rather than to a number on a marketing page. What defeats a rate limit is not how many addresses exist but how fast they turn over: the pool is rebuilt continuously, and roughly 40% of the exits a target saw yesterday are gone today. An IP-level ban therefore has a short horizon, and per-IP budgets reset along with the address. When a flow needs one IP to stay put, pin a session instead.
Plenty of sites decide what you are from the TLS handshake, before they have seen a single header. Anti-Detect Mode negotiates TLS and HTTP/2 with the fingerprint of a real Chrome, Firefox or Safari build, so the request looks like the browser it claims to be.
Some portals expose their records only through JavaScript search forms, postback navigation or session-bound pagination. Every plan includes cloud Chrome workers you drive over HTTP — navigate, fill forms, paginate, screenshot, export PDF. Same account, same traffic pool, same bill.
A retry bug at 3am can spend a month of traffic before anyone is awake. Give each proxy user its own daily ceiling and a bad job stops at its own limit instead of the account's. Datacenter and residential are capped separately, and both reset at 00:00 UTC.
Scrapers do not fail loudly. A site changes its markup or starts serving 403s, and the job keeps running — requesting, being refused, spending your allowance on responses you cannot use. We record the status the target returned for every request, chart your success rate next to your traffic, and let you set a floor. Fall below it and that proxy user is switched off and you are emailed. Your others keep running.
Drive a headless browser through any proxy and it will dutifully fetch webfonts, analytics beacons and ad tags you are going to throw away. Block them by pattern and they are refused before anything is fetched — no connection, no bytes, nothing billed.
Only responses from the target site count toward your success rate. Errors we return — traffic limits, blocked hosts, your own request filters — are never counted against you.
Point your HTTP client at one endpoint. No proxy lists to rotate, no browser farm to run, no SDK to install. Works with Python, Node.js, Go, Scrapy, Playwright and any HTTP client.
curl -x x.botproxy.net:8080 \
-U "user:password" \
https://example.comEach request rotates to a fresh IP, or pin a session to hold one. Anti-Detect Mode spoofs TLS fingerprints so requests are not rejected as obvious automation.
Proxy documentationFrom $50/month with 50 GB included. All locations on every plan. Volume tiers up to 1.2 TB.
Cloudflare-hardened targets with managed challenges. Consumer platforms behind login. Sneaker drops and ticket resale. Datacenter IPs with TLS spoofing will not reliably get through those, and no amount of fingerprint work changes that — you want residential, and we would rather say so than sell you a plan that fails.
Test your target above before you sign up. If it does not come back clean, we are not your vendor.