用自建多区域探针实测分流出口、DNS / WebRTC 泄漏、AI/流媒体可达性与延迟带宽,给出可追溯证据。
Cyber Metadata · IP 纯净度体检
判断当前出口 IP 是否像住宅、机房、VPN、代理或共享基础设施,给出可追溯证据和评分。
Cyber Metadata · IP 纯净度体检
判断当前出口 IP 是否像住宅、机房、VPN、代理或共享基础设施,给出可追溯证据和评分。
用自建多区域探针实测分流出口、DNS / WebRTC 泄漏、AI/流媒体可达性与延迟带宽,给出可追溯证据。
IP 纯净度检测判断一个出口 IP「像不像真实终端用户网络」:是住宅 / 移动宽带,还是机房、云主机、CDN 一类共享基础设施,或是 VPN / 代理 / Tor 出口。这决定了它在注册风控、支付、流媒体解锁和 AI 服务这些场景下容不容易被拦截。
每一项证据都在结果页单独列出(判定依据 / 证据链),可逐条核对来源与置信度,不是不可解释的黑盒分数。
综合多个自建信号:ASN 类型(机房/住宅/移动)、威胁情报命中密度、住宅代理特征库、反向 DNS、被动 DNS 历史等。每一项的加减分都单独列出,可在结果页的判定依据里逐条核对,不是黑盒分数。
不一定。纯净度衡量的是「像不像真实终端用户网络」,VPN/代理/Tor 是单独识别的布尔特征,两者会同时展示——一个住宅代理出口可能纯净度看起来不低,但仍会被标记为高风险。
支持。IPv4 / IPv6 都能查询,双栈出口(同时有 v4 和 v6 公网地址)会分别列出,并检测两者是否一致。
不够。纯净度只是静态画像,想知道分流出口是否一致、DNS/WebRTC 是否泄漏、AI 服务是否可达,运行完整深度检测。
An IP cleanliness check judges how much an exit IP looks like a real end-user network — residential / mobile broadband versus shared infrastructure such as datacenters, cloud instances and CDNs, or a VPN / proxy / Tor exit. That distinction drives how likely it is to be blocked by sign-up risk control, payments, streaming unblocking and AI services.
Every signal is listed individually on the result page (evidence chain) with its own source and confidence — not an unexplainable black-box score.
It combines several self-built signals: ASN type (datacenter / residential / mobile), threat-intel hit density, residential-proxy fingerprint databases, reverse DNS and passive-DNS history. Every contributing signal is listed on the result page's evidence chain — it is not a black-box number.
Not necessarily. Cleanliness measures how much an IP looks like a real end-user network; VPN/proxy/Tor are separate boolean signals shown alongside it — a residential-proxy exit can score reasonably clean while still being flagged high-risk.
Yes. Both IPv4 and IPv6 are supported; a dual-stack exit (public v4 and v6 both present) lists each address and checks whether the two are consistent.
No — it is a static profile only. To confirm split-exit consistency, DNS/WebRTC leaks and AI reachability, run the full deep check.