What a proxy does in scraping
A proxy sits between your scraper and the target site and forwards the request from its own IP address. The site sees the proxy's IP, not yours. That matters for two reasons: sites rate limit and block by IP, and many sites show different content depending on where a visitor appears to be.
Proxy pools are usually split by where their addresses come from. The two big groups are datacenter proxies and residential proxies. Neither is better in every case; they trade cost and speed for how ordinary the traffic looks.
Datacenter proxies
Datacenter proxies use IP addresses that belong to hosting providers. They run on servers with fast, stable connections, so they have low latency, high bandwidth and a low price per gigabyte.
The downside is reputation. IP ranges are publicly registered to their owners, and it is easy for a site to see that a request comes from a hosting network rather than a home or a phone. Ordinary visitors rarely browse from a datacenter, so many sites treat that traffic with suspicion: they show more challenges, apply stricter rate limits or refuse it outright. Whole ranges can be blocked at once.
Where they work well
- Public APIs and sites without bot protection.
- Large downloads where bandwidth cost dominates.
- Internal testing and monitoring of your own sites.
Residential proxies
Residential proxies route traffic through IP addresses that internet providers assign to home connections. To a target site, a request looks like it comes from an ordinary household. That makes them much less likely to be blocked on reputation alone, and it makes country targeting more realistic, since the address really belongs to a network in that country.
They cost more and are slower. Home connections have less bandwidth and more latency than servers, and individual addresses come and go. A good residential setup compensates by rotating to a new address when one fails.
Where they work well
- Sites with bot protection or strict rate limits per IP.
- Region-specific content such as prices, availability and search results per country.
- Any target where a block costs more than the extra proxy traffic.
Rotation and sticky sessions
Independent of the pool type, you choose between rotating and sticky addresses. Rotation gives every request, or every retry, a new IP. That spreads load and means a retry after a block comes from a fresh address. Sticky sessions keep the same IP for a while, which you need when a site ties state to a visitor: paginated results, a cart, or a multi-step flow.
In Siftzo, every attempt gets a new IP by default. Pass the same session value on several requests to keep one exit IP for about 10 minutes:
curl https://api.siftzo.com/v1/scrape \
-H "X-API-Key: $SIFTZO_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"url": "https://example.com/results?page=2",
"country": "de",
"session": "listing_run_7"
}'A session controls only the IP. Each request still gets a fresh browser profile, so cookies are not carried over; send the cookies you need as headers. Use the same country on every request in a session, and start a new session name if the current IP gets blocked, because retries within a session use the same address.
How Siftzo uses proxies
Every Siftzo request, on both the HTTP and the browser tier, leaves through a residential proxy. You do not configure proxies; you set country to an ISO 3166-1 alpha-2 code when location matters, and session when you need a sticky IP. Proxy traffic is included in the request price, with no separate bandwidth charge. See the pricing page for current prices.
To be clear about what is not offered: Siftzo does not currently have a datacenter pool, so you cannot choose a cheaper datacenter exit for easy targets. You can use the network through the Scraper API and the Browser API, or point your own client at it with proxy access, billed per MB.
The number of available addresses differs per country. Small countries can see more retries or proxy_unavailable errors, which, like other proxy errors, are not billed. The details are in the proxies and sessions docs.
Choosing for your project
A simple rule of thumb:
- If the target has no protection and you move a lot of data, datacenter proxies are the cheaper tool.
- If you see blocks, need country-accurate content or scrape many different sites, residential proxies save time.
- If a site needs JavaScript, the proxy is only part of the answer; see headless browser scraping.
Whatever the pool, proxies do not make it acceptable to ignore a site's terms or hammer it with requests. Keep request rates reasonable and read handling anti-bot challenges responsibly.
Try it on your own data
1,000 free credits every month, no card needed. Failed requests are never billed.
Questions
Does Siftzo offer datacenter proxies?
No. Siftzo currently sends every request through residential proxies and does not have a datacenter pool.
Can I use Siftzo's proxies with my own HTTP client?
Yes. With proxy access you point any HTTP client or browser at proxy.siftzo.com, sign in with an API key and pay per MB of traffic. See the proxy access docs.
How long does a sticky session keep the same IP?
About 10 minutes. Requests that pass the same session value in that time leave from the same exit IP. Sessions are scoped to your account.
Is proxy bandwidth billed separately?
No. Proxy traffic is included in the price of each request. See the pricing page for current prices.