Anthropic is becoming the focus of attention in the technology community when temporarily suspending Peter Steinberger's account, the creator of the OpenClaw open source project, before quickly restoring it in just a few hours.
The incident was announced by Steinberger on social network X, accompanied by a photo of an announcement from Anthropic stating that his account was locked due to "suspicious activity".
However, the ban did not last long. After the information spread and attracted many comments, his account was reopened.
In a public exchange, an engineer from Anthropic affirmed that the company "has never banned anyone just for using OpenClaw", and offered direct support.
Although it is not clear whether this is the reason for the account recovery or not, the incident has sparked much debate about how AI companies manage their ecosystems.
This development occurred shortly after Anthropic announced policy changes related to third-party tools.
Accordingly, AI assistant Claude's registration packages will no longer include support for tools like OpenClaw. Users who want to continue using it will have to pay a separate fee through the API, based on consumption level.
Anthropic explains that previous registration packages were not designed to handle the usage model of automated agent systems.
Systems like OpenClaw can consume a lot of resources due to their ability to run continuous inference loops, automatically retest tasks and connect to many other tools.
However, Steinberger disagrees with this explanation. He believes that the policy change move coincides with the time Anthropic added similar features to its own system, especially the remote agent coordination tool in internal services.
The developer also emphasized that he still complies with the new regulations and uses a valid API, but his account is still locked. This further made him express disappointment, while implying the difference in how companies approach the development community.
Some opinions online also raise the issue of Steinberger currently working at OpenAI and being an opponent of Anthropic.
However, he affirmed that his two roles are independent, one is to develop OpenClaw towards multi-platform compatibility, and the other is to participate in product orientation at OpenAI.
Notably, Steinberger said he is still testing Claude to ensure OpenClaw works stably with this platform, as it is still a popular choice for project users.
This incident shows the increasing tension between commercial AI models and the open source community.
As major platforms tighten controls and adjust price policies, third-party developers may face more barriers in maintaining compatibility and product development.
