Blog
How agents use computers, and how to give every agent, or every customer, a computer of its own.
- Meta's Muse gives every user a cloud computer. Here's how to do the same for your customers Muse runs each user's agent on its own virtual machine with a browser and stored credentials. How it works, and how to build it in a Next.js app with the Vercel AI SDK and Burrowbox.
- ChatGPT agent works on its own computer. Build the same with the OpenAI Responses API OpenAI's ChatGPT agent and Dots do their work on a virtual computer, with takeover for logins and approval before consequential actions. How to give your own agent that setup.
- Claude computer use needs a computer. Give it one through the Claude API's MCP connector Claude's computer use runs in an environment you provide. How that works, what Anthropic recommends, and how to connect Claude to a persistent Burrowbox machine with the Messages API.
- Project Mariner runs ten tasks at once on cloud VMs. Fan out your own agents with LangChain Google moved its browsing agent to virtual machines in the cloud so it can run tasks in parallel. How to do the same with a warm pool of Burrowbox machines and LangChain.
- Comet puts an assistant in your browser. Put one in a cloud browser that works on a schedule Perplexity's Comet builds an agent into the browser you use. How to run the same kind of browsing agent on a signed-in cloud browser, triggered by schedules and event webhooks.
- Manus gives every task its own cloud computer. Build it as a durable workflow on Vercel Manus runs each task in an isolated cloud VM with a browser, files and tools. How to build the same per-task computer with a Burrowbox warm pool and Vercel's Workflow SDK.