Claude Account Bans, Suspensions and Appeals

A user asks people who have used Claude long-term from mainland China without ever being banned to share their setup: VPS, VPN, residential IP, and registration or payment region.

Explains that a proxied browser still leaves the terminal on a direct connection, so running claude leaks the real IP and triggers an instant ban, plus three checks: proxy env vars, handshake test, and TUN takeover.

A reply laments that Japanese users can freely use Claude without worrying about account bans, hinting at ban pressure in other regions.

A user says their Claude Pro account has survived 1.5 years using a relatively clean VPN and a reseller top-up, believing Anthropic knows they are in China but both sides play along.

A poster says they heard Claude recently started unbanning previously banned accounts and asks for confirmation from anyone just restored.

The author argues that new users registering with dirty IPs and hopping between IPs start with a very low safety score, making bans inevitable and getting their devices flagged too.

After his main Claude account was banned and Codex degraded, the author moved two content pipelines to DeepSeek and was surprised by its search, no-setup skills and low cost.

A user tried four Google accounts to access Claude: one was banned right after login and two were banned seconds after KYC, including one bought on Xianyu.

A user asks whether installing Claude Code CLI on a Grok bot's cloud machine and never logging into the Claude account from phone or PC would basically prevent bans.

A user says a Claude account bought through a proxy registration service was banned within days, then pivots to selling pre-made accounts with an order link.

A seller of ready-made Claude accounts warns that even with a ban guarantee, buyers who skip the setup package must fix their own usage environment or keep getting banned, linking a tutorial and Telegram group.

A reply suggests paying for Claude with Apple gift cards to lower risk and get refunds if banned, adding that risk controls mainly target Max while Pro is safer.