Aperture GA: model tokens, MCP controls, Projects, and custom permissions
- Source
- tailscale.com
- Author
- Tailscale
- Date
Context
Tailscale's Aperture, an AI gateway for agents connected to a Tailscale network, gives those agents tools for working with devices on that network. Its general-availability announcement describes interfaces for adding machines and accessing them through SSH, so an can help configure a service without the user manually copying credentials between systems.
The access remains governed by network rules. Tailscale says users approve each machine Aperture adds and agent actions are logged. Projects can also share instructions, permitted tools, and network devices across related chats. These documented controls let users define the working environment an agent receives, rather than re-establishing it in every conversation.
- AI agent — An AI system that doesn't just answer once but works toward a goal in a loop — taking actions, reading the results, and deciding what to do next.
- guardrails — The checks around a model that block bad inputs and outputs — filters, validators, and permission rules the model itself can't override.
- MCP — The Model Context Protocol — an open standard that lets any AI assistant plug into any tool or data source without custom integration code.
- tool permissions — The rules governing which tools an agent may call and which need confirmation — the boundary between a mistake and an incident.
Checking sign-in…
Loading comments…