Home / Technology / Anthropic Cuts Off OpenClaw, Citing Capacity Issues
Anthropic Cuts Off OpenClaw, Citing Capacity Issues
4 Apr
Summary
- Anthropic ended support for OpenClaw subscriptions due to usage patterns.
- A recent source code leak for Claude Code was attributed to human error.
- Users affected by the change receive a one-time credit or can buy discounted bundles.
Anthropic has ceased support for the popular AI agent platform OpenClaw for its Claude subscriptions, a move effective Saturday at 12 p.m. PT. This decision stems from the platform's inability to accommodate the usage patterns of third-party tools, as outlined by Boris Cherny, head of Claude Code. "Our subscriptions weren't built for the usage patterns of these third-party tools," Cherny stated.
This change comes shortly after the source code for Anthropic's Claude Code command line interface application was leaked widely on GitHub, an incident Anthropic attributed to "human error." Cherny emphasized that capacity management is crucial for serving customers sustainably, and prioritizing direct customer usage is key to managing growth.
Affected subscribers will be provided a one-time credit equivalent to their monthly subscription cost. Users can still access OpenClaw with their Claude login by purchasing discounted usage bundles or by utilizing a Claude API key. This allows for continued, albeit modified, access to the platform.
Peter Steinberger, the creator of OpenClaw, voiced his disappointment with Anthropic's decision, noting that attempts to negotiate with Anthropic only delayed the change by a week. He also pointed out Anthropic's recent feature copying before implementing the lockout. Despite his criticism, Steinberger credited Cherny for attempting to soften the impact of the announcement and mentioned that upcoming OpenClaw updates would improve cache usage and reduce costs for API users.
Prior to these events, Anthropic had also begun enforcing stricter session limits on Claude, reducing token allowances during peak usage hours. In a separate development, Peter Steinberger was recently hired by OpenAI to lead the development of next-generation personal agents, with OpenAI stating its commitment to supporting OpenClaw as an open-source project. Analysts have suggested this acquisition by OpenAI was a missed opportunity for Anthropic, given OpenClaw's original integration with Claude.
