You guys, Anthropic just dropped a huge update to Claude Cowork that makes remote AI management insanely convenient. If you've used their coding agent at all this year — which launched in January after a capability expansion in March — now there is an official mobile layer for it. After updating your Claude app on Android or iOS you get a dedicated Cowork tab right in the sidebar to track what the model is doing on your computer back home, and I think that's exactly how they want you using it. One quick clarification before we go further: this isn't an autonomous agent running tasks directly on your phone — Claude won't be clicking apps for you on mobile — instead it lets you keep a window into the desktop session from anywhere. That is actually the right framing because what Anthropic really built was background tasking that no longer requires a constant stable internet connection to complete jobs, which is massive for workflow continuity when things get offline.
And check out the safety guardrails they've baked in — if Cowork needs permission on any specific step it fires a notification straight to your phone and you tap through; Anthropic literally says nothing ships until you review and approve it first. They also laid out their roadmap which is pure gold: they're planning to merge Cowork into the main chat interface so one conversation handles both natural language chatting AND computer tasks from a single pane, rolling out on web/desktop soon. On top of that they are pulling Projects and Artifacts together too — remember Projects lets you bundle files and chats around a single topic to maximize what Claude can actually hold in its context window at once while Artifacts is where the model builds mini apps and even games for you live. If you're a Max subscriber you get this update first, with rollout to other tiers expected in the coming weeks.
Source: https://www.engadget.com/2209495/now-you-can-direct-Anthropic-Claude-Cowork-AI-from-your-phone
And check out the safety guardrails they've baked in — if Cowork needs permission on any specific step it fires a notification straight to your phone and you tap through; Anthropic literally says nothing ships until you review and approve it first. They also laid out their roadmap which is pure gold: they're planning to merge Cowork into the main chat interface so one conversation handles both natural language chatting AND computer tasks from a single pane, rolling out on web/desktop soon. On top of that they are pulling Projects and Artifacts together too — remember Projects lets you bundle files and chats around a single topic to maximize what Claude can actually hold in its context window at once while Artifacts is where the model builds mini apps and even games for you live. If you're a Max subscriber you get this update first, with rollout to other tiers expected in the coming weeks.
Source: https://www.engadget.com/2209495/now-you-can-direct-Anthropic-Claude-Cowork-AI-from-your-phone