How to Check an IP Fraud Score Before You Rely on a Proxy
By the ProxyV4 Engineering Team · Updated:
To check an IP fraud score, find the proxy's real exit IP, then look it up in two or three reputation services such as IPQualityScore and Scamalytics, plus a blocklist check. Each service computes its own score from its own data, so the numbers will not match, and none of them tells you how a particular website will treat the IP. Use the results as one input, then test the IP on the job you actually need it for.
How to check an IP fraud score
A reputation lookup only means something for the address websites actually see. Connect through the proxy first and read the exit IP, then look that IP up. Write down each result with the date, because scores change as vendors update their data.
Manual lookups
- Exit IP: open the my IP tool in a browser that uses the proxy, or call
https://proxyv4.com/api/my-ipthrough it. - Fraud or risk score: paste the IP into the lookup page of a reputation vendor such as IPQualityScore or Scamalytics. Free lookups are usually rate-limited.
- Network type: most lookups also show the network owner and whether the vendor labels the IP as hosting, proxy, VPN or residential.
- Blocklists: run the IP through the free IP blacklist check to see whether public DNS blocklists list it.
Scripted lookups with an API
For more than a few IPs, use a vendor's API. The script below reads the exit IP through the proxy, then queries IPQualityScore's IP reputation API directly (not through the proxy) with your own key. Field names follow IPQualityScore's documentation; other vendors return different fields.
import requests
PROXY = "http://USER:[email protected]:PORT"
proxies = {"http": PROXY, "https": PROXY}
# 1. The address websites see when you use this proxy
exit_ip = requests.get("https://proxyv4.com/api/my-ip", proxies=proxies, timeout=10).json()["ip"]
print("Exit IP:", exit_ip)
# 2. One vendor's view of that address (needs your own API key)
r = requests.get(f"https://ipqualityscore.com/api/json/ip/YOUR_API_KEY/{exit_ip}", timeout=10)
r.raise_for_status()
data = r.json()
if not data.get("success"):
raise SystemExit(data.get("message"))
for field in ("fraud_score", "proxy", "vpn", "ISP", "ASN", "country_code"):
print(field, "=", data.get(field))Keep within the vendor's plan limits and terms, and store the results with a timestamp so you can compare later lookups against them.
What a proxy risk score is built from
Reputation vendors do not publish their formulas. What they describe publicly is the kind of data they collect: who owns the network the IP belongs to and whether it is a hosting, business or consumer network; whether the IP has been seen in abuse reports, spam traps or honeypots; whether it behaves like an open or public proxy; and what other sites in their network have reported. The weighting, the data sources and how long old events count are their own.
Why two vendors disagree
Each vendor sees different traffic, buys or builds different feeds and updates on its own schedule. One may score an IP as high risk because of a report another vendor never received. Scales differ too: IPQualityScore reports a 0 to 100 fraud score, Scamalytics shows its own score with a risk label, and other services return yes or no flags instead of a number. Comparing raw numbers across vendors is not meaningful.
A proxy or hosting label is a description
A datacenter IP is, by definition, on a hosting network, and many vendors label it as hosting or proxy for that reason alone. That label says what kind of network the IP is on. It does not mean the IP was used for abuse, and for scraping, API work or monitoring it usually makes no difference.
IP reputation check: blocklists versus fraud scores
Blocklists and fraud scores answer different questions. A blocklist is a public list with stated listing reasons, often about email abuse; a fraud score is a private estimate. The table shows how to read each kind of result.
| Result | Typical source | What it tells you | What to do |
|---|---|---|---|
| Listed on a DNS blocklist | Public DNSBLs such as Spamhaus | Someone reported mail or abuse from the IP or its range | Read the listing reason; most lists explain how delisting works |
| Hosting, proxy or VPN flag | Reputation vendors, geolocation databases | The type of network the IP is on | Expected for datacenter IPs; decide whether your target cares |
| High fraud score | One vendor's private model | That vendor's view, based on data you cannot see | Check other vendors and test on the real target |
| Location differs between lookups | Geolocation databases | Databases disagree, often on newly assigned IPs | Compare sources; report a clear mismatch to the IP owner |
Reading the result for your own job
A website may use one of these vendors, several, or none, and it does not tell you which or how it weighs them. That is why a low score elsewhere does not mean a site will accept an IP, and a high score does not mean it will refuse it. The useful test is the job itself: buy one IP, run the real workflow at normal volume for a few days, and watch for errors or challenges.
If a site you do not operate refuses datacenter traffic, that is its policy. Use its official API, or ask the site owner for access. Do not use proxies to get around a site's fraud checks, verification or account limits, and follow each platform's rules.
What ProxyV4 IPs look like to reputation services
ProxyV4 private datacenter IPv4 addresses sit on hosting networks, so expect many vendors to label them as hosting or proxy. Static ISP IPs are registered to consumer internet providers in the United States and Vietnam, and vendors may classify them differently, but no provider can promise how a third party scores an address. Each IP has one tenant for the whole term, so while you hold it, nobody else's traffic is added to its history.
Because scores vary, the honest way to buy is to order one IP, check it with the steps above, and test it on your real workload before you scale. Renewing before expiry keeps the same IP, so an address that works for you stays yours. Traffic is unmetered and every proxy supports HTTP(S) and SOCKS5 with username and password.
The refund policy covers technical failures, such as a proxy that cannot connect or shows the wrong location, credited to your account balance; read it for the exact scope before ordering. If you need mobile carrier IPs or a new IP on every request, ProxyV4 is not the right fit. Live per-IP rates are on the pricing page.
Habits that keep an IP's record readable
- One IP per workflow: when each IP does one job, a problem in its history is easy to trace.
- Moderate request rates: respect each site's rate limits and back off when it asks you to, rather than pushing volume through one address.
- Credentials kept private: ProxyV4 proxies require a username and password; do not share lines between teams or put them in public repositories.
- Re-check after incidents: if a site starts refusing an IP, run the lookups again and compare them with your first results before changing anything.
Frequently asked questions
What is a good IP fraud score?
There is no universal cutoff. Each vendor uses its own scale and publishes its own guidance, and websites rarely say which vendor, if any, they use, so test the IP on your real workload.
Why do IPQualityScore and Scamalytics show different scores for the same IP?
They collect different data, weigh it differently and update on different schedules. Compare each vendor's result with its own earlier results, not with another vendor's number.
Can I lower an IP's fraud score?
Not directly. Stop any activity that could be reported as abuse and give it time; for a blocklist entry, follow that list's published delisting process.
Does a proxy flag mean the IP is bad?
No. A hosting or proxy flag describes the network type. It matters only if the site you use treats that network type differently.
Will I keep the same IP if I renew?
Yes. Renewing a ProxyV4 proxy before it expires keeps the same address, and you can renew all proxies or a selection.
Test reputation yourself before scaling
Order one private IPv4 or static ISP IP, run the lookups and your real workload on it, then add more. Same IP on renewal, unmetered traffic.
See live prices