Claude Account Bans, Suspensions and Appeals

Anthropic updated its usage policy to prohibit sustained, pointless mistreatment of its models starting November 12, with offenders first warned and then rate-limited or suspended.

A user whose Claude account was suspended emailed Anthropic's safety team, then filed an appeal through the website, and got access back in about six days, sharing the original email template.

A user reports that after two months of testing, both of their Claude accounts remain unbanned, calling the situation fairly stable for now.

A fourth roundup of Claude tips covers forcing a system proxy on the web client to block tracking probes, an appeal template for banned accounts, and backing up AI memory locally to GitHub.

A user says they will research how to use Claude without being banned and report back in a month, reflecting anxiety over account bans.

The poster says they use Claude via a VPN with a Singapore native residential IP, bought no dedicated IP, and was never banned, calling most anti-ban guides marketing hype.

A user asks whether remotely connecting to the system running Claude, such as a Mac mini kept at a friend's home overseas, would still be detected and lead to a ban.

One item in a financial morning brief says Anthropic updated its usage policy to prohibit abusive or cruel behavior toward its AI models.

A commenter argues that complaints about ChatGPT being slow, Claude account bans and when Argon launches all boil down to whether the money is worth it, adding that only users who have been banned make the trending list.

A user says that to avoid being banned they moved abroad, avoided the UTC+8 time zone, switched to English prompts and paid with a local girlfriend's bank card.

The poster believes their Claude account has not been banned mainly because they are a light Pro subscriber who rarely uses up the weekly quota.

The author had Codex talk directly to Grok Bot to verify whether the network setup for running Claude on their computer was configured correctly, confirming no IP leaks but noting this does not guarantee the account won't be banned.