Anthropic launches Claude Duet, a proactive personal agent

Started by Tracey, Oct 25, 2026, 10:55 AM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

Topic: Anthropic launches Claude Duet, a proactive personal agent   Views(Read 83 times)

Tracey

Anthropic has announced a new personal agent called Claude Duet. Engadget's Igor Bonifacic reports that it is Anthropic's answer to Meta's Muse and OpenAI's Dots. As always, it is worth noting that Anthropic makes Claude, so this is a company announcing its own product

Claude Duet is a proactive agent. It sends a short daily briefing of action items from connected apps, and can handle tasks such as scheduling meetings, building dashboards, making forecasts and prototyping web apps. It checks with the user before taking any action

Duet runs on Opus 5.5 in its own virtual machine, so it can work in the background, and it has standalone apps for iOS and Android. Access is currently through a waitlist for US Claude subscribers only, as a research preview

Anthropic says 20 dollar a month Pro users will get it when it becomes generally available, whereas OpenAI's Dots needs a Pro plan costing 100 dollars or more. Duet shares usage limits with Claude Code. Anthropic also says it never sees passwords or sign in tokens directly

Proactive agents are the new battleground for AI companies, as the Meta Muse privacy problem showed this week. Asking before acting is a sensible design choice. Pricing it for ordinary subscribers could help it spread quickly, if it works well

Would you let an AI agent manage your schedule and tasks? And does asking before acting make you more comfortable with it?


Eastern Nathan

US only is a shame for UK users. Hopefully it expands soon. Most launches start in America. Wider rollouts usually follow within months

Crossing

Meta's Muse privacy problem shows how careful these products need to be. Sharing personal data with agents carries risk

Design choices matter. Trust will decide which agent people choose. Security design matters as much as features. Anthropic seems to know that

Megan97

Daily briefings sound useful. Pulling action items from email and calendars saves time. I would try it

WWEDanny96

Sharing limits with Claude Code might frustrate heavy users. Developers already hit those limits

We will see how it balances. Usage limits are a common frustration with AI tools

Balancing them across products is tricky. Developers will watch closely

Pale Connor

Checking before acting is the right approach. Agents doing things without permission worries people

This builds trust. Permission before action should be standard for agents. Agents are becoming the next big battleground

Dom_41

Twenty dollars versus one hundred dollars is a big price difference. Anthropic is clearly competing on access. That could win users. Pricing will matter as much as features. Many people will not pay 100 dollars a month