Anthropic Drops OpenClaw Support After Code Leak Sparks Controversy

Support for the OpenClaw AI Agent was discontinued by Anthropic from Cloud Subscriptions. This came shortly after the source code for the Cloud Code command-line interface tool from Anthropic, but not including the models themselves were leaked on GitHub. The leak appears to be the result of an internal mistake on the part of Anthropic. In fact, the company has openly stated that the leak was due to a "human error."

Boris Cherny, head of Cloud Code, said in a post on X (formerly Twitter) that starting Saturday at 12 noon PT, Cloud subscriptions will no longer support third-party tools like OpenClaw. "Starting tomorrow at 12 noon PT, Cloud subscriptions will no longer cover usage on third-party tools like OpenClaw." He added, "We're working hard to meet the growing demand for Cloud, and our subscriptions weren't designed for the usage patterns of these third-party tools. Capacity is a resource we manage thoughtfully, and we're prioritizing customers who are using our products and APIs."

Subscribers will receive a one-time credit equal to the price of their monthly plan. If you need more, you can now purchase discounted usage bundles. To request a full refund, see the link in your email yesterday. We want to manage our growth in a way that ensures we can continue to offer our customers a stable service over the long term. This change is a step in that direction.

They informed users that they can "still use these tools with your Cloud login via Additional Usage Bundles (which are now available at a discounted rate) or with a Cloud API Key."

In response, OpenClaw creator Peter Steinberger said, "I woke up this morning and my mentions were full of this stuff. @davemorin and I both tried to convince Anthropic, but we could only delay it for a week. It's funny how the timing matches. First they copy some popular features into their closed system and then push open source out."

In another post, he said that while what Anthropic has done is sad, he credits Cherny for softening the blow. He wrote, "While I think what Anthropic is doing is a tragedy for the entire ecosystem, I would like to give Boris credit for doing their best to mitigate the negative impact. Today's update will include some improvements for better cache utilization, reducing costs for API users."

Prior to this news, Anthropic had tightened Cloud session limits to five hours of use during working hours (5 a.m. to 11 a.m. PT / 8 a.m. to 2 p.m. ET). This reduced the number of tokens you could send during those sessions.

Peter Steinberger was recently hired by OpenAI. Released in November 2025, iOpenClaw is an open-source AI agent that created a stir in the developer world and raised concerns among enterprise security teams. In February, Steinberger joined OpenAI.