Coding agents have made significant advancements, becoming capable of generating and reviewing code with impressive complexity. As interest in these tools grows among engineers and executives, organizations face a challenge: how to enable secure AI usage without creating barriers that developers will circumvent. Tailscale's response is Aperture, an AI gateway currently in alpha that aims to provide visibility into coding agent usage while minimizing disruption for developers. Aperture integrates with various coding tools, including Claude Code and Codex, and simplifies the process of managing API keys. It leverages Tailscale’s identity framework, allowing organizations to bypass the need for multiple API keys across different environments. Users can set up Aperture with minimal effort, and once configured, it streamlines access to various AI models and providers by associating API usage with user and machine identities. This feature not only facilitates easier adoption of AI tools but also enhances security and monitoring. Visibility is another key function of Aperture. It helps organizations analyze AI adoption and usage efficiency, which is crucial for managing costs associated with token usage. By consolidating this data, it enables teams to identify unusual or unauthorized activity in real-time. Aperture can export usage logs for integration with other systems, enhancing overall security and compliance. Tailscale plans to expand Aperture beyond coding agents to cover other AI applications within organizations. The platform is designed to be extensible, allowing companies to add additional security features. While it’s currently in alpha and may have bugs, Tailscale encourages feedback from users to refine the product. Interested users can sign up for a free trial, but those with larger teams can join a waitlist for dedicated instances.