IPure IP purity check

IP purity report

103.152.226.189

80/ 100

Fair

confidence: high

Usable; platforms with strict risk controls may occasionally show a CAPTCHA.

Based on signals observed from public and partner data sources at 2026-10-24 17:53 UTC. This describes IP-level risk only; platforms' internal risk data is not visible to us.

Profile

What this IP is

Country / region
Hong Kong · Central and Western District · Hong Kong
IP type
Business broadband
Native / broadcast
Native IP
Network
AS32167 XM Cloud
Reverse DNS
Anti8-ddos.xdpcloud.com
Devices seen in subnet
no data

Suitability

What it is good for

Each use case weighs the evidence differently; these are not one overall score.

AI services

ChatGPT · Claude · Gemini

—

Region not supported

ChatGPT / Claude / Gemini are not offered in this region and it cannot connect directly; a clean IP in a supported region is needed

Least tolerant of data center and proxy exits: more likely to get degraded service, verification prompts or account limits. Spam history does not matter here

Social sign-up

X · Discord · Telegram · Reddit

88

Excellent fit

No notable obstacle found for this use case

Sign-up is the strictest checkpoint: data center / proxy exits are more likely to be asked for phone verification or refused, and IPs with bulk-registration history are blocked far more often

Streaming / short video

TikTok · Netflix · YouTube

95

Excellent fit

No notable obstacle found for this use case

Native IP and low sharing matter most: a broadcast IP reads as a region mismatch, and shared exits get throttled

Gaming platforms

Steam · PSN · Epic · Nintendo

96

Excellent fit

No notable obstacle found for this use case

Region locks, regional pricing and gifting all follow the IP's location, so broadcast IPs tend to be treated as cross-region; data center exits carry a high risk of restrictions

Cross-border e-commerce

Amazon · eBay · Shopee

89

Excellent fit

No notable obstacle found for this use case

Linked-account store closures are driven first by shared IPs, and only then by data center ranges

Email sending

SMTP · marketing email

62

Borderline

Main issue: Blocklist hit: DroneBL

Blocklists make or break deliverability, while sending from a data center IP is normal and is not penalised

Evidence

Why it got this score

  • −12

    Blocklist hit: DroneBL

    Listed by DroneBL

    What it means
    This IP is on a public blocklist. Most lists record addresses that sent spam or took part in scanning or attacks.
    Which uses care
    Cares most: Email sending; Barely looks at it: AI services, Social sign-up, Streaming / short video, Gaming platforms
    Why it is weighted this way
    A single aggregate list is moderate evidence, so about a dozen points; the email scenario more than doubles it because receiving servers query these lists directly.
    Can it be fixed
    Every list has a removal process, provided the cause is gone; some delist automatically after a quiet period.
  • −10

    1 open ports

    Includes admin ports 22: clear evidence of a server rather than home broadband

    What it means
    Scanners see ports open on this IP. With admin ports such as 22 or 3389 it is almost certainly a server, not a home router.
    Which uses care
    Barely looks at it: AI services
    Why it is weighted this way
    Mainly corroboration that "this is a server", which the datacenter item already charged for, so only a few points here; AI-type scenarios weight it at 0.3 to avoid charging the same fact three times.
    Can it be fixed
    Close or restrict ports that do not need public access (firewall, non-default port, fail2ban); scanners update within days, then re-check. One of the few items you can change yourself.
  • −3

    IP type: Business broadband

    Business broadband is broadly trusted, though a small share are proxy exits

    What it means
    A business broadband address: generally trusted, though some corporate exits are used as proxies.
    Which uses care
    Cares most: AI services, Social sign-up, Gaming platforms, Cross-border e-commerce; Barely looks at it: Email sending
    Why it is weighted this way
    A small deduction meaning "slightly less certain than residential".
    Can it be fixed
    An inherent attribute of the IP or its range; nothing a user can change short of a different IP.
  • +5

    Native IP

    Registered and used in the same country, allocated directly by a local carrier

    What it means
    Registered and used in the same country, allocated directly by a local carrier: a "native IP".
    Which uses care
    Cares most: Streaming / short video, Gaming platforms; Barely looks at it: Email sending
    Why it is weighted this way
    A bonus: region-based checks will not trip.
    Can it be fixed
    An inherent attribute of the IP or its range; nothing a user can change short of a different IP.

Reputation

Blocklists and abuse history

Listed on 1 of 8 blocklists checked: DroneBL.

  • abuseipdb abuse confidence 0%

Open ports observed: 22

Limits

What an IP check cannot tell you

Each platform's internal risk labels
Platforms keep their own IP reputation data and ban records and do not publish them. No hit in public sources does not mean a platform has nothing on this IP.
Whether the IP is truly exclusive
Sharing is only the number of devices observed in the same subnet. It cannot show who else is using this exit right now, or what they are doing.
Which accounts this IP has been tied to
Account-to-IP links exist only inside each platform. Accounts previously banned on this IP are invisible to us.
Device, browser and behaviour
Risk systems look at IP + device fingerprint + behavioural rhythm together. A clean IP with an odd fingerprint or behaviour still gets blocked.
Whether the account profile matches the IP
When sign-up country, phone number, payment method, language or time zone disagree with the IP's location, platforms verify further. That is unrelated to IP purity.

Sources

Data sources and responses

  • rdap
    169ms
  • team-cymru
    63ms
  • ip-api.com
    94ms
  • proxycheck.io
    156ms
  • abuseipdb
    620ms
  • cloud-ranges
    10ms
  • tor-exits
    10ms
  • rdns
    1181ms
  • shodan-internetdb
    181ms
  • dnsbl
    205ms

Machine-readable versions of this report: JSON · plain text · 中文报告