On 20 October 2026 Anthropic added two features to the Claude Platform for companies that serve many end customers through one API account: user profiles that say which end customer a request is for, and organization webhooks that report when Anthropic restricts or bans one of those profiles.

What Anthropic Shipped

The Claude Platform release notes for 20 October say: "We've launched user profiles, so you can tell Anthropic which of your end customers each API request is for. If you operate a platform or gateway, or resell API access, a safety enforcement action can then apply to one end customer instead of your whole organization." The SDK and the ant CLI support the new /v1/user_profiles endpoints.

The same entry adds organization webhooks to the Admin API, in beta behind the webhooks-2026-10-07 header. They cover two event types: "one for when Anthropic restricts or bans one of your organization's user profiles, and one for when the restriction or ban ends." Access goes through the customer's Anthropic account team. The same day, inference hooks left beta for Claude Enterprise organizations.

What It Means For Companies Running Agents

A company that runs agents for many customers on one model account has had one blast radius: misuse by a single customer could restrict the whole account. With user profiles, the provider can act on one customer, but only for requests that carry a profile. Tagging every request, including scheduled and background runs, is the step that makes the narrower action possible.

The webhooks turn a restriction into an event a system can handle. A company can pause that customer's agents and queued jobs, keep the logs for that profile, and assign a person to the review, while other customers keep running. It can resume when the end-of-restriction event arrives and its own review is closed.

What This Does Not Settle

  • The release notes do not say what conduct triggers a restriction or a ban.
  • They do not describe an appeal process or how long a restriction lasts.
  • The webhooks are in beta and need a request to the account team.

Related: The agent overreach containment playbook covers pausing one agent without stopping the rest. See also our notes on the Anthropic usage policy update.