Urgent.News

What's breaking now, across thousands of outlets.

AI

Why We Run Our AI Agent on Your Mac Instead of a Cloud Computer

A lot of personal AI agents now come with their own computer: a virtual machine in the cloud with a browser, a file system and a model driving it. It's a clean design. The agent gets a sandbox it fully controls, and nothing it does can touch your machine. The catch shows up on the first real task. Your files are on your laptop. Your logins live in your browser. Your mail, calendar and the odd…

Rather than running AI agents on remote cloud computers, personal AI agents like Nico run directly on your Mac. This approach allows the agent to work with your local files, logins, and applications without the need to upload documents or sign into accounts again remotely. The design separates the agent into two parts: the front agent, which converses with you and handles immediate tasks, and the worker, which performs the actual work, plans, reads files, interacts with apps and browsers, and checks results.

The worker uses an open-source agent harness called pi to manage model interactions, subscriptions, and session handling. Nico customizes pi by adding its own tools, compaction summaries, and risk-level classification for each step. The agent can perform tasks ranging from reading documents to composing emails, with riskier actions like payments requiring manual approval.

Running the agent locally ensures seamless integration with your existing applications and login credentials, while the ability to use various models through your own accounts eliminates the need for additional AI subscriptions. Nico offers a free plan for Apple silicon Macs running macOS 15.4 or later, with an optional Plus plan for enhanced features.

Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

Read the original at dev.to →

More in AI

I asked an AI to find bugs and got 0. Changing one line of the prompt found 5.

You often hear that AI bug checks can't be trusted because the AI just says "no issues found." I'm not an engineer. I build apps together with AI. While doing that, I saw the number of bugs found go from 0 to 5 with the same code and the same model — the only change was one line of the prompt.

  • AI bug checks often claim 0 issues found, despite presence of bugs
  • Modifying prompt by adding one line increased bug detection from 0 to 5
  • Fable excels at persistence, finding 7 bugs in 12 apps with only 1 attempt

More from Wednesday 4 November →