Claude Account Bans, Suspensions and Appeals

A satirical song with lyrics, music and an MV made by Claude Code about a Pro subscription being banned right after payment.

A user complains that Claude is banned, Kimi's quota is used up and DeepSeek is out of service, saying they can no longer program without AI.

A user who says they have had over 30 Claude accounts banned now runs three stably, crediting a dedicated, consistent IP and App Store payment with an Oregon tax-free address, and calls timezone changes, fingerprint browsers and account warming placebos.

The user advises keeping projects local, writing a handoff document and local git commit after each task, and periodically backing up Claude or Codex folders and conversation logs so work can continue on another tool after a ban.

A user says a Claude Pro account they had just topped up for 150 RMB was banned right after discussing one question, and plans to switch to Codex and try Grok Bot.

After two paid Claude accounts were banned within three days, the user paid via US App Store and US PayPal, used a Japanese residential IP and a Taiwan timezone, and the account has now lasted nine days.

A user says that beyond changing IPs they also have to talk to Claude in broken English, and asks for advice on how to keep an account from being banned.

The author says ChatGPT Plus, Claude Pro and a dedicated clean IP cost nearly 300-400 yuan a month, yet Claude still banned the account without warning.

After comparing the official terms with popular anti-ban guides on X, the author says only three rules hold up: use your own account, keep the exit IP and payment region fixed, and avoid bulk script runs.

A user argues that Claude bans on mainland accounts are unsurprising because several domestic platforms have drained the service's resources.

A user says they stayed with Claude from Opus 4.8 through Opus 5.5 while others flocked to Codex, then had their account banned right as Claude surged.

A user pushes back on the claim that cutting off Chinese users would doom Anthropic, saying only insufficient model and tool capability can kill it.