Anthropic caused confusion after updating its Claude Code documentation to state that using OAuth tokens from consumer plans in third-party products would violate its terms of service. The change alarmed users of personal AI agents like OpenClaw and NanoClaw, which rely on Claude Pro and Max subscriptions rather than pay-as-you-go API keys. Anthropic subsequently clarified that the update was a routine documentation cleanup and that nothing is changing about how customers use their accounts with the Agent SDK. The company stated it will not cancel accounts but advised developers building commercial services on top of the SDK to use API keys instead.

Anthropic: You can still use your Claude accounts to run OpenClaw, NanoClaw and Co. - The New Stack