Anthropic stops Claude subscriptions from covering third-party AI agents like OpenClaw

Anthropic is ending a loophole that let some users run Claude subscriptions through third-party AI agents such as OpenClaw, according to VentureBeat and Anthropic support materials.

The reported change means Claude Pro and Max subscriptions will no longer cover usage through outside agent tools. Users who want to keep using Claude in that way will need separate extra usage or API-based billing.

Anthropic's help center says extra usage is billed at standard API rates after a plan's included limits are reached. The company also says Claude Pro has fixed usage limits and no option to buy additional or unlimited messages on the plan itself.

For Claude for Work plans, Anthropic says extra usage can be enabled and is also billed at API pricing once included limits are exhausted.

The policy shift matters for power users and agent builders who had been routing Claude through third-party tools. It also shows Anthropic drawing a sharper line between consumer subscriptions and agent-style usage that resembles API traffic.

VentureBeat reported the change on April 4, 2026, and Anthropic's support pages provide the billing framework behind it. The company has not publicly detailed how many users or tools are affected beyond OpenClaw and similar services.