Anthropic Blocks Subscription Access to OpenClaw and Third-Party Tools Amid Service Strain

Summary

Anthropic has prohibited subscription-based users from accessing third-party agentic tools like OpenClaw that interface with Claude, citing engineering constraints and the need to optimize system capacity. The restriction takes effect immediately and represents an effort to manage surging demand for Claude services, which has strained the company's infrastructure. Boris Cherny, head of Claude Code, explained that the company's systems are optimized for specific workloads and that the measure is necessary to serve as many users as possible with the most advanced models. The move follows a partial outage on Claude.ai on April 6 that affected login functionality, conversations, and voice mode, demonstrating the pressure Anthropic's systems are currently experiencing. Subscription plans offer significantly better pricing than API alternatives—some users reported getting up to 36x value—making them attractive workarounds for cost-conscious developers. The policy aligns with similar enforcement actions taken by Google regarding its Gemini CLI and OAuth authentication access.

• The policy mirrors Google's approach to protecting API access and terms of service compliance across its AI platforms

Editorial Opinion

Anthropic's move to block third-party subscription access reveals the tension between maximizing accessibility and maintaining operational viability at scale. While the company's engineering constraints are legitimate, this restriction effectively punishes cost-conscious developers who found legitimate workarounds—a common pattern across cloud platforms. The real issue isn't OpenClaw's existence, but rather Anthropic's pricing model, which creates such a dramatic arbitrage opportunity that limiting it becomes necessary for the business. As AI companies race toward profitability and public markets, expect similar restrictions to become standard practice.