Anthropic ends Claude Code usage on OpenClaw

Anthropic says subscriptions for Claude (including Claude Code) will stop covering usage on third-party tools like OpenClaw starting April 4 at 12pm PT. The company is framing the move as a capacity-management step, effectively limiting how far paid Claude plans can be consumed through external integrations.

For teams relying on OpenClaw with a Claude Code subscription, this changes the cost and availability of everyday workflows. Instead of having subscription coverage extend into that third-party environment, usage will no longer be covered once the cutoff hits.

Third-party agent tooling can create unpredictable demand spikes, since developers may run multiple tasks, long-running agent loops, or higher-volume workloads than a provider expects from native product interfaces. Anthropic's decision suggests it's prioritizing more predictable utilization of model capacity.

It also underlines a broader pattern in AI platforms: vendors increasingly draw boundaries around what "subscription" includes, especially as usage shifts from direct API calls to agentic tools and marketplaces. After the cutoff, OpenClaw-based usage may require new billing arrangements or different tooling strategies, depending on what Anthropic and OpenClaw make available.

The next practical question is how OpenClaw will handle the transition for existing users -- whether it offers alternative plans, changes routing to other supported models, or updates pricing for Claude-backed workloads. Users will likely want to verify how their current integration billings will behave after April 4 at 12pm PT.