AWS Machine Learning Blog→ original

AWS launched a self-hosted gateway for enterprise governance of Claude Code and Desktop

AWS introduced Claude apps gateway for AWS, a self-hosted gateway that gives organizations a single point of control for access, costs, and corporate policies for Claude Code and Claude Desktop. The gateway is deployed on the organization's own infrastructure and works with Amazon Bedrock, allowing IT teams to control developers' use of AI tools without data leaving the corporate perimeter.

AI-processed from AWS Machine Learning Blog; edited by Hamidun News
AWS launched a self-hosted gateway for enterprise governance of Claude Code and Desktop
Source: AWS Machine Learning Blog. Collage: Hamidun News.
◐ Listen to article

AWS launched a self-hosted gateway for corporate management of Claude Code and Desktop

On July 9, 2026, AWS introduced Claude apps gateway for AWS — a self-hosted management gateway that gives organizations a single point of control over access, spending, and policies when working with Claude Code and Claude Desktop in a corporate environment.

What Claude apps gateway can do

The gateway acts as an intermediary between employees using Claude tools and the company's cloud infrastructure. AWS describes three key areas of control.

Access — IT administrators determine which employees can use Claude Code and Claude Desktop, in what capacity, and under what conditions. This eliminates "shadow AI" scenarios where developers connect to tools directly, bypassing corporate security policies.

Cost — the gateway provides visibility into AI tool spending. Organizations can set consumption limits, track expenses by teams and projects, and obtain analytics for budget planning.

Policies — companies establish rules at the gateway level: restrict transmission of certain data categories, log requests for auditing, apply corporate security standards and regulatory requirements.

Why self-hosted matters

Self-hosted deployment is a fundamental architectural decision for the enterprise market. All requests to Claude Code and Claude Desktop pass through the organization's own infrastructure, not through public cloud endpoints.

For companies with strict compliance requirements — financial sector, healthcare, government agencies — this is critical. Developers transmit sensitive data to AI tools: proprietary code, internal documentation, client information. A self-hosted gateway ensures this data remains within the corporate perimeter.

The gateway works with Amazon Bedrock — AWS's managed service for accessing foundational AI models, including Anthropic's Claude family. Companies already using Bedrock can integrate the gateway into existing infrastructure without building a separate stack. Claude Platform on AWS provides an additional integration layer.

Context: AI tools have become corporate infrastructure

Claude Code has established a strong position among AI assistants for developers. As its use scales across large teams — when dozens and hundreds of engineers use the tools — manageability becomes a top priority: who has access, how much does it cost, what data is being sent outside.

A similar pattern is already implemented by competitors: Microsoft offers enterprise management for GitHub Copilot, Google offers administrative mechanisms for Gemini Code Assist. Claude apps gateway fills an analogous gap in the AWS and Anthropic ecosystem.

The announcement reflects an important trend: AI developer tools are ceasing to be personal applications and are becoming corporate infrastructure with corresponding requirements for manageability, security, and economics.

What this means

Claude apps gateway is a practical answer to the main barrier to corporate AI adoption: the lack of centralized control. Companies with AWS infrastructure can now make Claude Code a standard tool for their entire engineering team without sacrificing security and budget control requirements.

ZK
Hamidun News
AI news without noise. Daily editorial selection from 400+ sources. A product by Zhemal Khamidun, Head of AI at Alpina Digital.

Need AI working inside your business — not just in your newsfeed?

I build production AI for companies — custom CRM, internal tools, autonomous agents, workflow automation. Owned by you, shaped to your process, no per-seat tax. Built by Zhemal Khamidun, CPO of AlpinaGPT (AI platform, 6,000+ users).

What do you think?
Loading comments…