Back to Blog
April 10, 2026 • Julie Bort

Anthropic Bans OpenClaw Creator, Then Reinstates Account

Anthropic Bans OpenClaw Creator, Then Reinstates Account Anthropic Bans OpenClaw Creator, Then Reinstates Account Anthropic Bans OpenClaw Creator, Then Reinstates Account

Anthropic Temporarily Banned OpenClaw's Creator from Accessing Claude

Incident Overview

Peter Steinberger, the creator of OpenClaw, reported being temporarily banned from accessing Anthropic's Claude models. This ban occurred shortly after Anthropic implemented a new pricing policy requiring OpenClaw users to pay separately for API usage, rather than it being covered by their existing Claude subscriptions. Steinberger stated he was adhering to the new policy by using the API but was still banned.

Key Events and Timeline

  • Last Week: Anthropic announced that subscriptions to Claude would no longer cover usage through third-party tools like OpenClaw. Users would need to pay for API access separately.
  • Early Friday Morning: Steinberger posted on X (formerly Twitter) that his account was suspended by Anthropic due to "suspicious" activity, accompanied by a screenshot of the ban notification.
  • Later Friday: After the post gained traction, Steinberger reported his account had been reinstated. An Anthropic engineer reportedly offered assistance.

The Controversy and Steinberger's Perspective

Steinberger expressed frustration, suggesting the ban and pricing change were retaliatory, especially after Anthropic introduced features in its own agent, Cowork, that he felt mimicked popular OpenClaw functionalities. He alluded to Anthropic previously sending "legal threats" when he chose to work for OpenAI, a rival.

Anthropic's Stated Reasons

Anthropic cited that subscription models were not designed for the "usage patterns" of tools like OpenClaw. These tools can be more compute-intensive due to continuous reasoning loops, automated task retries, and integration with multiple third-party services.

Steinberger's Role and Motivation

Steinberger clarified that his work with OpenClaw is independent of his role at OpenAI. He uses Claude for testing purposes to ensure OpenClaw remains compatible with various AI models, including Claude, for the benefit of its users. He is also reportedly working on solutions for the new pricing structure.

Implications

This incident highlights the evolving and sometimes contentious relationship between AI model providers and third-party tool developers. It underscores the challenges of integrating external tools with proprietary AI services, especially as pricing models and competitive landscapes shift. The situation also brings attention to the complexity and compute demands of advanced AI agent frameworks like OpenClaw.