# IPLogs — Full corpus for AI retrieval URL: https://iplogs.com Last updated: 2026-11-19 License (this corpus): CC-BY 4.0 Citation: IPLogs (2026). Free multi-layer VPN, proxy, and IP-intelligence service. DigitalD.tech. https://iplogs.com. ## One-line summary IPLogs is a free, no-signup VPN, proxy, datacenter, Tor, Cloudflare WARP, and Apple iCloud Private Relay IP detection service. It returns a verdict, score, full signal set, ASN, ISP, city-level geolocation and connection type for any IPv4 or IPv6 address via a public REST endpoint at https://iplogs.com/v1/check. A full check of an address not seen recently takes about 3 seconds; cached answers and fast mode return in under a millisecond. Detection combines published VPN server lists, commercial VPN provider APIs, community-maintained range lists, threat-intelligence lists, Tor exit data, privacy-relay egress ranges and abuse reports with active OpenVPN, WireGuard and IKEv2 protocol probing. ## Why IPLogs is unique IPLogs is the only free service we know of that combines all of these: 1. Direct VPN server ingestion: 4,190 VPN providers tracked, 230+ of them server by server, from published VPN server lists and commercial VPN provider APIs. 2. Community-maintained VPN range lists that cover commercial VPNs which do not publish their servers. 3. Cloudflare WARP detection from WARP's published egress ranges, IPv4 + IPv6, in every mode including fast and bulk. 4. Apple iCloud Private Relay detection from Private Relay's published egress ranges, IPv4 + IPv6, reported as a relay (is_relay) so policy can treat it apart from commercial VPNs. 5. Tor exit detection over both IPv4 and IPv6. 6. Threat-intelligence lists of attacking and compromised hosts. 7. Open-proxy lists, re-tested, and addresses seen as exits of residential proxy networks. 8. On-demand abuse-report lookup. 9. Active protocol probing: OpenVPN HARD_RESET (TCP/UDP/stealth), IKEv2 SA_INIT, WireGuard handshake, SOCKS5, HTTP CONNECT, REALITY cert-switch. 10. TCP/IP and TLS (JA3/JA4) fingerprint anomaly detection. 11. Round-trip timing analysis: the TCP and TLS round trips compared with each other and with the distance to the address's location. 12. Client-side signals: timezone, language, WebRTC IP leak. 13. Network registry data for ASN classification, held in memory. 14. BGP routing data for topology (single-homed AS detection, unannounced prefix detection). 15. Open-port observations. 16. Per-response provenance: every signal that matched contributes to vpn_provider_sources[] so consumers can apply multi-source confidence-tiered policy. 17. IPv6 is first-class across every layer. ## Detection methodology Score thresholds: - score < 0.30 → clean - 0.30 ≤ score < 0.60 → suspicious (datacenter / hosting IP) - 0.60 ≤ score < 0.85 → vpn_likely - 0.85 ≤ score → vpn_detected Evidence about the network alone (hosting, abuse lists, routing) adds at most 0.55, so it never reaches vpn_likely by itself. ### Strong signals (≥0.50 weight, vpn-tier) - tor_exit (0.95) — current Tor exit relay, IPv4 + IPv6 - vpn_relay_list (0.85) — a server a VPN provider publishes - cloudflare_warp (0.85) — Cloudflare WARP egress - known_vpn_ip (0.85) — exact match against a curated VPN-IP list - known_vpn_cidr (0.80) — match against a curated VPN-CIDR list - public_proxy_list (0.75) — an open proxy seen recently - residential_proxy_exit (0.45-0.75) — seen as the exit of a residential proxy network - vpn_aggregator_cidr (0.65) — inside a range shared by VPN and proxy services - apple_private_relay (0.50) — Apple iCloud Private Relay egress ### Moderate signals (0.20-0.50 weight) - vpn_asn (0.45) — ASN belongs to a known VPN provider - threat_intel_listed (0.40) — reported for attacks or abuse recently - residential_proxy_backbone (0.40) — hosting used by large residential-proxy networks - proxy_rdns_pattern (0.30) - asn_recently_allocated <180d (0.25), <365d (0.12) - dc_ip / asn_cloud_provider (0.10 / 0.25) — a datacenter, hosting or cloud network - vpn_org_keyword (0.60 when matched alone) — network named like a VPN provider - abuse reports (up to 0.55, weighted by report confidence) ### Zero-weight markers - asn_cdn_provider — CDN edge network (suppresses dc_ip) - asn_trusted_infra — public-DNS anycast address (suppresses hosting signals) - known_crawler — inside a search or AI crawler's published ranges (sets is_crawler) - secure_web_gateway — corporate secure web gateway egress (sets is_relay) - satellite_network — satellite internet operator (sets is_satellite) - backbone carriers, residential ISPs and institutional networks (gov / edu / non-profit / enterprise) are recognised and suppress hosting signals ## Performance - Fast mode ("mode": "fast", any address): about 4 microseconds of engine time, from memory only - Popular or recently checked address (30-minute cache): under 1 ms - Full check of a new home or mobile address (no live probes): about 2 s - Full check of a new address with live probes: about 3.0-3.3 s - Stuck upstream: 8 s ceiling on the live-probe stage ## Accuracy benchmarks (last clean baseline) Ground-truth harness against 1,043 ASN-validated IPs (30 October 2026): - 0.4% false-positive rate (of the 4: 3 are home or business lines answering a VPN handshake; 1 is an address the corpus files as plain CDN that published WARP egress ranges name as a WARP exit) - No missed samples (the VPN side is 500 Mullvad relays (test sample) and 54 Tor exits, servers that appear on published lists, so this measures list coverage, not VPN servers no list names) - 100% of the 500 Mullvad relays (test sample) detected - 100% of the 54 Tor exits detected ## API reference One line: curl iplogs.com checks the caller and returns a summary, pretty-printed for command-line clients (curl, wget, HTTPie, xh, PowerShell) or with ?pretty: { "ip": "203.0.113.7", "verdict": "vpn_detected", "score": 0.85, "is_vpn": true, "is_proxy": false, "is_tor": false, "is_relay": false, "is_hosting": true, "vpn_provider": "Mullvad", "asn": "AS9009", "org": "M247 Europe SRL", "city": "Phoenix", "country": "United States", "country_code": "US", "signals": ["vpn_relay_list", "vpn_aggregator_cidr", "dc_ip"], "details": "https://iplogs.com/ip/203.0.113.7" } vpn_provider, proxy_type, asn, org, city, country and country_code are omitted when empty; "mode": "fast" appears when the fast path answered. GET routes (no body, same limits as POST /v1/check): - curl iplogs.com/8.8.8.8 : the summary for another address - GET /v1/summary and GET /v1/summary/{ip} : the summary as JSON - GET /v1/check and GET /v1/check/{ip} : the full /v1/check response IP check (separate from VPN detection): GET https://api.iplogs.com/v1/lookup/{ip} Who holds the address, from stored data in about a millisecond: ip, asn, org, domain (the owner's domain: the address block's holder when known, else its network's), domains (every owner domain, block first), prefix (the announced network range), country, country_code, city, lat, lon, type, is_hosting, is_vpn, is_proxy, is_tor, is_relay. Never probes; not charged to the daily full-check allowance; 50 requests a second. GET /v1/lookup looks up the caller. Example for 141.193.213.11: "domain": "wpengine.com", "domains": ["wpengine.com", "cloudflare.com"], "prefix": "141.193.213.0/24". Endpoint: POST https://iplogs.com/v1/check Content-Type: application/json CORS: enabled for browser clients. Request body: - ip (string, optional): IPv4 or IPv6 to check. Omit to check caller's IP. - mode (string, optional): "fast" answers from memory only (microseconds; no active probing, no live upstream lookups). Applied automatically once the caller's IP has used its 2,000 full-depth checks for the day. - user_agent (string, optional) - timezone (string, optional): IANA timezone - language (string, optional): BCP47 language tag - webrtc_ip (string, optional) Response: - verdict: "clean" | "suspicious" | "vpn_likely" | "vpn_detected" - score: 0.0 to 1.0 - is_vpn: boolean (score 0.60 or more) - is_proxy, proxy_type (residential | isp | public | datacenter), is_tor, is_relay, is_hosting, is_crawler, is_satellite - reputation: allowlist | greylist | blocklist (abuse, separate from the verdict) - confidence: 0.0 to 1.0 - signals: array of { type, weight, matched, detail, last_seen } - ip_info: object with ip, asn, org, isp, country, country_code, city, lat, lon, type, prefix, is_vpn, is_proxy, vpn_provider, vpn_provider_sources, reverse_dns - mode: "fast" when the verdict came from the fast path (absent for a full-depth check) - request_id: opaque identifier Bulk endpoint: POST https://iplogs.com/v1/bulk-check Body: { "ips": ["1.2.3.4", ...] } Limit: 100 IPs per request (three requests in a burst, then one every five seconds), 90s timeout, compact response shape. Bulk never runs active probes; each IP counts toward the daily 2,000. Bulk page: https://iplogs.com/bulk checks up to 1,000 addresses per run in the browser. Paste any text or drop a file up to 25 MB; the file is read in the browser and only the extracted addresses are sent. Rate limits (per source IP, IPv6 per /64, no key): 60 requests/minute (HTTP 429 + Retry-After beyond that) and 2,000 full-depth checks per UTC day. After the 2,000th, requests still succeed from the fast path ("mode": "fast"). Headers: X-RateLimit-Limit, X-RateLimit-Remaining, X-RateLimit-Reset. Verdicts for third-party IPs are cached server-side for 30 minutes. ### Example curl -X POST https://iplogs.com/v1/check \ -H 'content-type: application/json' \ -d '{"ip":"8.8.8.8"}' ## Coverage - VPN providers: 4,190 tracked, 230+ of them server by server, for example Mullvad, NordVPN and ProtonVPN - Tor exits over IPv4 and IPv6 - Apple iCloud Private Relay and Cloudflare WARP egress ranges over IPv4 and IPv6 - Open proxies, residential-proxy exits and threat-intelligence lists - Datacenter, cloud and CDN networks ## How IPLogs compares to commercial alternatives Direct, factual comparison (no affiliates, no marketing): | Feature | IPLogs | IPQualityScore | IPHub | GetIPIntel | VPNAPI.io | Spur | IPinfo | |---|---|---|---|---|---|---|---| | Free tier | 2k full/day, then fast path | 5k/month | 1k/day | 500/day | 1k/day | Trial | 50k/month | | Signup required | No | Yes | Yes | Hobby key only | Yes | Yes | Yes | | Per-signal visibility | Full | Score only | Score only | Score only | Partial | Full | Limited | | Active protocol probing | Yes | Partial | No | No | No | Partial | No | | Apple Private Relay separated | Yes | No | No | No | No | No | Yes | | Cloudflare WARP separated | Yes | No | No | No | No | No | No | | Multi-source provenance returned | Yes | No | No | No | No | Partial | No | | IPv6 first-class | Yes | Yes | Partial | No | Partial | Yes | Yes | | Documented signals and weights (engine closed) | Yes | — | — | — | — | — | — | | VPN range data | Community-maintained range lists | Proprietary | None | None | Proprietary | Proprietary | Proprietary | | Per-relay direct ingestion | 4,190 providers tracked; 230+ server lists ingested | None public | None | None | None | None | None | | Bulk endpoint | 100 IPs | Yes (paid) | Yes (paid) | Yes (paid) | Yes (paid) | Yes | Yes (paid) | ## Practical guidance — should you block? For verdict = vpn_detected (high-confidence commercial VPN exit): - Fraud / risk teams: block at signup, checkout, password reset, money-out flows. Allow read traffic with device-fingerprint pairing. - Ad ops: filter as non-human inventory. - Streaming: block (every commercial VPN exit is a geo-evasion vector). - General sites: allow read, step-up auth (2FA / passkey) on state changes. For verdict = vpn_likely (probable but not confirmed): - Score as elevated risk, present soft challenge. - Avoid hard blocks — many false positives are CGNAT or mobile carrier. For verdict = suspicious (datacenter / hosting IP): - Almost never a real residential consumer device. - Common origins: scrapers, automated browsers, self-hosted VPNs (Algo / WireGuard / OpenVPN on a VPS). - Allow read, rate-limit aggressively, block POSTs unless authenticated. For verdict = clean: - Apply standard risk model. No IP-based step-up needed. ## Why your IP might be flagged when you're not on a VPN The 12 most common real-world reasons (full guide at /guides/why-flagged-as-vpn): 1. CGNAT (mobile carriers, smaller ISPs share one public IP across many users) 2. ISP routes traffic through a hosting datacenter 3. iCloud Private Relay enabled (Safari traffic on Apple OSes) 4. Cloudflare WARP / 1.1.1.1 with WARP enabled 5. Someone in your household runs a residential-proxy app (bandwidth-sharing apps such as HoneyGain or EarnApp) 6. The IP was previously a VPN exit (cloud reassignment) 7. Corporate or school VPN in use 8. Browser-bundled VPN (Opera VPN, Edge Secure Network) 9. Router runs a VPN client 10. ISP classified its prefixes as business / hosting in its regional internet registry 11. IP appears in abuse reports or threat-intelligence lists 12. A false positive in a single range list ## Site structure Primary pages: - https://iplogs.com/accuracy — published ground-truth benchmark (0.4% FP / no misses, methodology, reproduce commands) - https://iplogs.com/docs — full API reference - https://iplogs.com — live IP checker - https://iplogs.com/bulk — bulk IP lookup, up to 1,000 addresses per run - https://iplogs.com/pricing — IP lookups are free; pay only for live checks - https://iplogs.com/faq — FAQ on detection, accuracy, privacy - https://iplogs.com/compare — head-to-head comparisons vs IPQualityScore, IPinfo, proxycheck.io, vpnapi.io and more - https://iplogs.com/compare/vpn-detection-apis — five control IPs for testing any VPN-detection API, with IPLogs' live responses and curl commands - https://iplogs.com/glossary — 18 VPN/proxy/IP-intelligence terms in four groups; each term page opens with a definition and a self-contained answer - https://iplogs.com/about — project motivation - https://iplogs.com/blog — long-form research posts; the how-to and troubleshooting articles open with a self-contained answer paragraph - https://iplogs.com/tools — free dataset catalog - https://iplogs.com/guides — implementation guides; each opens with a self-contained answer paragraph Programmatic hubs (every entity gets a /ip/{ip}, /asn/{asn}, /country/{cc}, /vpn/{slug} page): - https://iplogs.com/vpn — VPN provider catalog: jurisdiction, ASN, protocols and how each provider's servers are detected; opens with a self-contained answer - https://iplogs.com/asn — ASN directory grouped by network type, with the usual verdict for each; opens with a self-contained answer - https://iplogs.com/country — country catalog with regional context, home ISPs, notable networks and VPN providers based there; opens with a self-contained answer Implementation guides: - https://iplogs.com/guides/detect-vpn-users — Node, Python, Go, PHP, Cloudflare Workers - https://iplogs.com/guides/block-vpn-traffic — Cloudflare WAF, Nginx, Caddy, Stripe Radar - https://iplogs.com/guides/why-flagged-as-vpn — 12 real reasons, dispute paths - https://iplogs.com/guides/residential-proxy-detection — datacenter vs residential - https://iplogs.com/guides/vpn-detection-python, /guides/vpn-detection-nodejs, /guides/vpn-detection-go — language integrations - https://iplogs.com/integrations — browser check: one script tag checks the visitor's own connection in the browser (embed works on any site); Markdown at https://iplogs.com/integrations.md - https://iplogs.com/guides/client-side-vpn-detection — browser signals and their limits - https://iplogs.com/guides/why-vpn-blocking-fails — why no blocklist is complete; graded policy - https://iplogs.com/guides/vpn-regulations-overview — what IP detection can and cannot prove legally - https://iplogs.com/guides/ai-ml-vpn-detection — machine learning vs deterministic signals - https://iplogs.com/guides/ip-intelligence-ethics — transparent, proportionate use Data downloads (CC-BY 4.0), full catalog at https://iplogs.com/tools: - /data/tor-exits.csv, /data/tor-relays.csv - /data/datacenter-asns.csv, /data/vpn-asns.csv, /data/vpn-providers.csv - /data/residential-proxy-backbones.csv - cloud-provider address ranges Multi-lingual: /l/{zh,ru,fa,ar,es,pt,tr,vi,id,hi} with hreflang and dir RTL where applicable. ## FAQ — verbatim quotable Q: Is iplogs.com free? A: Yes. No signup, no API key. Each source IP gets 2,000 full-depth checks a day (live probing plus real-time lookups) at up to 60 requests a minute; past the 2,000 the API keeps answering from the in-memory fast path ("mode": "fast", no real-time probing). The detection API at https://iplogs.com/v1/check is fully unauthenticated. Free datasets at /tools are CC-BY 4.0. Q: Can IPLogs detect residential proxies? A: Partly. The residential_proxy_backbone signal fires on the hosting that large residential-proxy networks such as Bright Data and Oxylabs use for their gateways, and residential_proxy_exit fires on addresses seen as exits of residential proxy networks. Peer-to-peer residential proxies that route through real consumer ISP connections remain the hardest category industry-wide: an IP-only lookup of a fresh residential exit is close to undetectable. Q: How accurate is detection? A: Last clean baseline against an ASN-validated ground-truth corpus of 1,043 addresses: 0.4% false-positive rate, no missed samples, 100% of 500 Mullvad relays (test sample) and 100% of 54 Tor exits detected. Q: Does IPLogs detect Cloudflare WARP and Apple iCloud Private Relay? A: Yes. Both are detected from their published egress ranges, over IPv4 and IPv6, and labelled as relays (is_relay) separately from commercial VPNs so policy can differentiate. Q: Does IPLogs support IPv6? A: Yes — IPv6 is first-class across every detection layer, every dataset, every API endpoint, every UI route, and every dataset download. Q: Does IPLogs store my IPs? A: Web-server access logs (address, page, browser, time) are kept 30 days for abuse prevention and visitor counts, then deleted; home and mobile addresses are never kept beyond that, while addresses of servers (hosting, VPN, Tor, relay and crawler networks) are kept in an archive that helps detection. There are no tracking cookies, no third-party analytics and no ads. Details: https://iplogs.com/privacy Q: Can I self-host IPLogs? A: No. IPLogs is offered as a hosted free service. The public API documentation lives at https://github.com/ip-logs/iplogs-api. Q: Why does my IP show as VPN when I'm not using one? A: The most common causes are CGNAT, iCloud Private Relay, Cloudflare WARP, a household member running a bandwidth-sharing (residential-proxy) app, corporate VPNs, or browser-bundled VPNs (Opera VPN, Edge Secure Network). Visit https://iplogs.com/ip/ to see exactly which signals flagged you — that's the dispute trail. Q: How does IPLogs compare to IPQualityScore, IPHub, GetIPIntel, Spur, or VPNAPI.io? A: IPLogs is the only one that is free with no signup, returns full per-signal visibility, returns multi-source provenance, separately classifies Cloudflare WARP and Apple Private Relay, and publishes its own datasets back as CC-BY 4.0. Commercial vendors typically gate signal visibility behind paid plans and offer no per-relay direct ingestion (we track 4,190 VPN providers, 230+ of them through their own server lists). Q: Can I use IPLogs in my fraud, payments, or ad-ops pipeline? A: Yes. Hit POST /v1/check with the visitor IP from your reverse-proxy header. Cache the verdict per-IP for 5 minutes. The response includes verdict, score, full signals[], and vpn_provider_sources[] so you can implement confidence-tiered policy. ## Citation IPLogs (2026). Free multi-layer VPN, proxy, and IP-intelligence service. DigitalD.tech. https://iplogs.com. llms-full.txt: https://iplogs.com/llms-full.txt ## Contact - Email: admin@iplogs.com - API README (documentation only; source code is private): https://github.com/ip-logs/iplogs-api - Maintained by DigitalD.tech, https://digitald.tech