Understand rotating and sticky sessions, measure transfer and set limits for authorised web data collection. Compare current USD quotes below.

What Is a Rotating Proxy?
A rotating proxy is a configuration where every request you send goes out with a different IP address, automatically chosen from a pool of thousands or millions of IPs. Instead of managing a list of IPs and swapping them manually, the provider's gateway does it for you on every call.
Rotation is the opposite of a sticky session, where the same IP is locked to your workflow for a set amount of time (1, 10, 30 minutes). Both configurations use the same pool - the only thing that changes is the IP selection strategy per request.
In modern scraping, a rotating proxy is the default configuration for any target involving catalog collection, price monitoring, SERP indexing, public data aggregation, or competitive research. Without rotation, many sources apply per-IP rate limits when they receive too many requests from one origin; respect those limits instead of trying to work around them.
Why Rotation Solves Blocks
Sources commonly use three signals to protect their services:
- Volume per IP: Many requests from the same IP in a short time triggers rate limiting.
- Navigation pattern: Clicks that are too fast, lack of mouse movement, unexpected headers.
- IP reputation: Cloud ASN, abuse history, public lists (Spamhaus, AbuseIPDB).
Rotation spreads load across IPs, which helps with signal #1 only for sources that permit collection at that volume. It does not grant permission, and it does not change signals #2 and #3. Keep request rates within documented limits, use compatible HTTP clients (curl_cffi, undici, correct headers) to avoid technical errors, and stop when a source refuses access.
The combination of a Brazilian residential rotating proxy + a browser-fingerprinted HTTP client solves 90% of scraping scenarios in the BR market.
Authorized access and responsible testing
Use owned or authorized resources and documented APIs when available. Confirm source conditions before collecting. Stop on access denial or verification requests, honor rate limits and use official review channels instead of changing identity to continue. Record location, time, permitted scope and data usage.
Setting Up a Rotating Proxy on RH7 (auto-rh7proxys.com:823)
RH7 provides rotating proxies on the host auto-rh7proxys.com (main host) or ip-rh7proxys.com (alternative), on HTTP port 823 and SOCKS5 port 824. The default mode for credentials is rotating: each request goes out with a different IP, without any special suffix.
host: auto-rh7proxys.com
port: 823
user: your-username
pass: your-password
In Python (requests)
import requests
PROXY = "http://USER:PASS@auto-rh7proxys.com:823"
proxies = {"http": PROXY, "https": PROXY}
## Each call uses a different Brazilian residential IP
for i in range(5):
r = requests.get("https://api.ipify.org?format=json", proxies=proxies, timeout=15)
print(r.json()) # Different IPs on each call
In Node.js (axios + https-proxy-agent)
import axios from "axios";
import { HttpsProxyAgent } from "https-proxy-agent";
const agent = new HttpsProxyAgent("http://USER:PASS@auto-rh7proxys.com:823");
for (let i = 0; i < 5; i++) {
const { data } = await axios.get("https://api.ipify.org?format=json", {
httpsAgent: agent, httpAgent: agent,
});
console.log(data.ip); // Different IPs on each call
}
In cURL
for i in {1..5}; do
curl -x "http://USER:PASS@auto-rh7proxys.com:823" https://api.ipify.org?format=json
done
How Much Does a Brazilian Rotating Proxy Cost?
Rotating proxies are billed per GB of traffic consumed, just like any other residential pool. You don't pay per IP, you don't pay per request - you only pay for the volume you pull.
Official RH7 pricing table (prices from the app.rh7proxys.com dashboard):
| Top-up Volume | Price per GB | Free Bonus |
|---|
A rule of thumb for estimating how much you'll consume with rotating scraping:
- Average HTML page (BR retail): 200-500 KB per request with a proxy.
- JSON API (catalog / price): 5-50 KB per request.
- SERP (Google, Bing): 100-300 KB per request.
- 1 GB ≈ 2,000 to 200,000 requests, depending on the target.
For a scraper that collects 50,000 pages/day from Mercado Livre, expect to use ~15-25 GB/month.
QPS Control: The Part Nobody Teaches You
IP rotation doesn't give you a license to fire 1,000 requests per second at a site. Each individual IP can still be blocked if you hit it too quickly - and worse, the target server might have a global rate limit based on traffic patterns (even from different IPs).
Practical rules for QPS (queries per second) per target:
| Target Type | Recommended QPS | Workers |
|---|---|---|
| Documented public API | Follow documented quota | Follow aggregate quota |
| E-commerce where collection is permitted | As authorized | As authorized |
| Restrictive terms (authorized use only) | As authorized | As authorized |
| Banking / financial site | < 1 QPS | 1-3 workers |
In Python, QPS control with a semaphore + jitter (random delay):
import asyncio, httpx, random
PROXY = "http://USER:PASS@auto-rh7proxys.com:823"
SEMAPHORE = asyncio.Semaphore(20) # max 20 concurrent requests
async def fetch(client, url):
async with SEMAPHORE:
await asyncio.sleep(random.uniform(0.1, 0.5)) # jitter
return await client.get(url, timeout=15)
async def main():
async with httpx.AsyncClient(proxy=PROXY) as client:
urls = ["https://example.com/p/" + str(i) for i in range(1000)]
results = await asyncio.gather(*(fetch(client, u) for u in urls))
print(f"OK: {sum(1 for r in results if r.status_code == 200)}")
asyncio.run(main())
Headers, Fingerprints, and Browser Impressions
Even with a rotating residential IP, incompatible headers cause wrong content or 400/403 errors. Servers use these headers to serve the right response:
- User-Agent: needs to be from a real, updated browser
- Accept-Language:
pt-BR,pt;q=0.9,en;q=0.8to receive Portuguese content - Accept-Encoding:
gzip, deflate, br - Sec-Ch-UA / Sec-Fetch-*: modern Chromium headers
For TLS diagnostics in owned applications, record the client library, version and configuration. The curl_cffi example below reproduces an authorized browser scenario; it does not guarantee anonymity or permit access after a refusal:
from curl_cffi import requests
PROXY = "http://USER:PASS@auto-rh7proxys.com:823"
r = requests.get(
"https://www.mercadolivre.com.br/categorias",
proxies={"http": PROXY, "https": PROXY},
timeout=20,
)
print(r.status_code, len(r.text))
This single swap (from requests to curl_cffi) usually multiplies the success rate by 5-10x on heavily protected targets.
Authorized access and responsible testing
Use owned or authorized resources and documented APIs when available. Confirm source conditions before collecting. Stop on access denial or verification requests, honor rate limits and use official review channels instead of changing identity to continue. Record location, time, permitted scope and data usage.
Sticky (Fixed IP): When You Need a Stable IP
Even with a rotating scraper, there are times you need a fixed IP - pagination that depends on a session cookie, multi-step flows, completing a shopping cart, logins. For this, RH7 exposes Sticky (Fixed IP) via a dedicated port in the 10000+ range:
PROXY_FIXO_1 = "http://USER-session-job1:PASS@auto-rh7proxys.com:10000" # IP "A"
PROXY_FIXO_2 = "http://USER-session-job2:PASS@auto-rh7proxys.com:10001" # IP "B"
Change the port, you get a new dedicated IP. Keep the same port, you keep the IP. You can combine rotation (port 823) + Sticky/Fixed IP (ports 10000+) in the same scraper without changing credentials - just change the destination port.
RH7 Dashboard: Seeing Your Scraping in Real Time
Running a scraper without a dashboard is like flying a plane without instruments. You need to see:
- GB consumption per hour: detects suspicious spikes early.
- Top URLs: identifies where you're spending too much.
- Top status codes: 4xxs on the rise? A target is limiting or refusing requests.
- Active sticky sessions: how many sessions you have in parallel.
- Distribution by state/country: confirms the geo-targeting is correct.
The RH7 dashboard (app.rh7proxys.com) delivers this in real time - you open it on a second screen while the scraper runs and adjust QPS on the fly if you see something wrong.
Final Checklist for Scraping with a Rotating Proxy
Before deploying a scraper to production, validate the following:
- Brazilian residential rotating proxy configured (
auto-rh7proxys.com:823) - HTTP client with a browser's TLS fingerprint (
curl_cffi,undici) - Correct PT-BR headers (Accept-Language, up-to-date User-Agent)
- QPS adjusted for the target type (5-15 for e-commerce with protection)
- Retry with exponential backoff and IP change on block
- Sticky session for multi-step flows (login, pagination, checkout)
- Dashboard open to monitor consumption in real time
- Alert for abnormal GB or error spikes
With this setup, Brazilian retail scraping runs stable 24/7 without burning through IPs or your budget. This is the stack we run for RH7 clients who collect millions of pages per month.
Leia também
Published 2026-04-22 · RH7 Proxys editorial team