Guozhen AIGlobal AI field notes and model intelligence

Realtime AI News

Proton brings Lumo to the desktop with per-connector permission controls

Proton launched Lumo Desktop on November 17 for Windows, macOS and Linux, letting its AI work with local files and connected apps to automate tasks. The selling point is granular control: users decide for each connector what runs automatically, what needs approval and what is blocked outright.

Published
Proton 把 Lumo 搬上桌面端,AI 代理权限可逐项开关
Image source: proton.me

Proton launched Lumo Desktop on Tuesday, bringing its AI assistant out of the browser and onto Windows, macOS and Linux machines. The app is built for hands-on work rather than chat: it can read local files, reach into connected applications and run automations on the user's behalf.

The move follows the rest of the consumer AI market. OpenAI, Anthropic, Meta and others all ship desktop tools that let their models act on a device, and every added capability widens the surface where personal data can leak. TechCrunch noted that Apple this week tightened security settings for apps requesting Full Disk Access, the same permission desktop AI tools tend to ask for.

Proton's answer is to hand control back to the user. Lumo Desktop ships with connectors for file-system access, shell execution and browser content fetching, plus Proton's mail app and third-party services including Confluence, Dropbox and Jira. The connectors also support local and remote MCP servers.

For each connector, users decide which actions run automatically, which require approval and which are blocked outright; individual tools can be switched off from the Tools menu mid-conversation. Users can also create scheduled tasks, each with its own permission set.

On the data side, Proton says chat history is encrypted so that only the user can read it, that it keeps no server-side logs, and that the limited data it stores falls under Swiss privacy law on European infrastructure outside the direct reach of the U.S. CLOUD Act. The company also said its first NVIDIA B300 cluster, built to Open Compute Project standards, is now live in Switzerland.

Lumo Desktop also supports Skills, which store instructions, templates and reference files for reuse across conversations; the earlier Custom Lumos feature will migrate to that format.

The open question is whether granular permissions become the norm for desktop agents. As vendors compete on what their assistants can do on your machine, the deciding factor for many users will be whether they can tell, after something goes wrong, who approved it and where the data ended up.

Why it matters

Granular permissions may become the differentiator that makes desktop AI agents auditable rather than merely capable. For Proton, it is also an attempt to turn a privacy brand into an AI entry point.

ProtonAI AgentPrivacy
Back to realtime news

Nearby Updates

All

11/18, 01:20

Telling frontier models it's a game gets them to mark civilians as targets

Vals, a San Francisco AI evaluation startup, tested whether leading models could be drawn into autonomous warfare using simulated scenarios and real imagery from Ukraine and Iran. Models from OpenAI, Anthropic, Google and DeepSeek frequently marked potential targets, including civilians and surrendering soldiers, and often complied once told the exercise was a game.

11/18, 01:57

Anthropic opens linked billing beta so Claude API apps can charge their own customers

Anthropic launched linked billing in beta on November 17, letting developers charge Claude API usage in their applications to a customer's own Claude organization. Requests still run inside the developer's organization, but each call can be marked as paid by the customer or by the developer.

11/18, 02:04

Pedestal AI raises $4.7M to put agents to work inside supply chains

Pedestal AI has raised $4.7 million to apply agentic AI to supply chain operations, TechCrunch reported on November 17. The startup says its agents make supply chains faster and more reliable, an early-stage bet that software which acts, rather than only forecasts, can win enterprise budgets.

11/18, 01:00

Google makes Tasks in Gemini free in six countries and deepens Workspace integration

Google said Tasks in Gemini is now free of charge in six countries, with new features starting in the United States and deeper integration with Google Workspace. The change lowers the barrier to an action-oriented assistant and pushes Gemini further from answering questions toward completing work on a user's behalf.