Anthropic has launched the Claude apps gateway, a management plane for running Claude Code across Am...
The AMW Read
This is an incremental product update for a known player in the foundation model segment, but it meaningfully advances enterprise governance for coding agents, impacting the segment's adoption dynamics.
Anthropic has launched the Claude apps gateway, a management plane for running Claude Code across Amazon Bedrock and Google Cloud in enterprise environments. The gateway centralizes authentication, policy enforcement, usage tracking, and cost management, announced on June 29, 2026 (US time). This product targets organizations that have adopted Claude Code but previously had to issue credentials per developer and manage policies separately across cloud platforms.
The gateway integrates with enterprise identity providers like Google Workspace, Microsoft Entra ID, and Okta via OpenID Connect, enabling short-lived sessions without storing long-term credentials on developer endpoints. Administrators can define policies once on a server, which are automatically applied to all requests. It also includes usage metering that sends data via OpenTelemetry Protocol (OTLP) to enterprise-owned observability platforms, and supports routing requests to Claude API, Amazon Bedrock, or Google Cloud with failover capability.
For builders and enterprises, this tool addresses a key governance gap: previously, managing Claude Code at scale across multiple cloud environments required bespoke infrastructure for identity and policy control. The gateway promises centralized cost management with limits set by organization, group, or user, and ensures that non-API data and usage metrics are not transmitted to Anthropic unless enterprises explicitly choose Claude API. Anthropic is also open-sourcing the reference implementation, allowing enterprises to deploy their own gateway, which could set a precedent for how coding agent governance tools are standardized.
