Anthropic Purges Claude Accounts Using VPNs in Unsupported Regions

A simple commercial VPN is no longer enough to reach Claude from officially unsupported regions. Users outside sanctioned territories—most prominently in Hong Kong and Russia—are reporting sudden permanent bans from Anthropic, losing their chat histories alongside active Pro and Max subscriptions.

Reports surfacing on community hubs such as Reddit describe accounts being suspended for "suspicious signals" despite uninterrupted connections over paid VPN tunnels. Claude remains officially unavailable in mainland China, Hong Kong, and Russia, among other territories. While geographical blocks were previously trivial to bypass by tunneling network traffic, Anthropic appears to have tightened its anti-fraud systems.

A standard VPN only reroutes traffic and masks external IP addresses at the network layer. According to technical accounts of the crackdown, Anthropic looks beyond raw IP locations by combining passive and active device fingerprinting. The detection relies on three main giveaways:

• Hosting ASN flags: Datacenter IP addresses associated with popular commercial VPN providers are easily cataloged and flagged as anomalous by anti-fraud heuristics.
• Environment mismatches: Connections masquerading as residential endpoints in Singapore or the United States frequently leak conflicting system clocks, localized OS languages, or regional keyboard layouts.
• WebRTC leaks: Unpatched browsers routinely expose local and true public IP addresses through WebRTC APIs, entirely bypassing the VPN tunnel.

The sudden enforcement leaves paying customers without recourse or refunds, as using a circumvention tool directly violates the service's terms of access. Whether users transition to antidetect browsers and residential proxies or simply abandon the platform, Anthropic has made its position clear: access from outside supported borders is an active target, not an overlooked loophole.

Source: habr