---
title: "Infrastructure"
description: "What an AI agent needs from the browser it drives: where the browser runs, how the model talks to it over MCP and what each step costs in tokens."
canonical: "https://scalebrowser.net/blog/category/infrastructure"
---

> ## Documentation Index
> Fetch the complete documentation index at: https://scalebrowser.net/llms.txt
> Use this file to discover all available pages before exploring further.

# Infrastructure

> What an AI agent needs from the browser it drives: where the browser runs, how the model talks to it over MCP and what each step costs in tokens.

An agent's browser is infrastructure, like its model or its job queue. These articles compare the ways to give an agent one, on your own machine or in a vendor's cloud, and measure what the connection between model and browser costs per step. The numbers come from our own bench and name the tool versions they were taken with.

For how Scalebrowser connects to an agent, the [documentation](/docs) is the reference.