Anthropic launches Claude Duet as a research-preview agent for professionals
The agent, built on Claude Opus 5.5, is available on iOS, Android, and the web, with a waitlist for US Pro and Max subscribers.
1 source · cross-referenced
- Anthropic announced Claude Duet, an AI agent designed for professionals and small business owners, launching as a research preview.
Anthropic announced Claude Duet, an AI agent designed for professionals and small business owners, launching as a research preview. The agent is available as a standalone app on iOS and Android and through Claude on the web. Access is currently limited to a waitlist, with eligibility restricted to US-based Pro or Max subscribers.
Claude Duet is built on Claude Opus 5.5 and integrates memory, skills, and connected apps from the broader Claude ecosystem. It provides a daily brief of important items across connected apps and alerts users only when intervention is needed.
The agent can perform administrative tasks such as scheduling meetings, as well as more complex workflows like building dashboards with multiple tools, creating financial or sales forecasts, prototyping, and building web apps. It operates within a virtual machine with a browser to complete tasks.
Before taking sensitive actions such as sending payments or deleting data, Claude Duet requests user approval. It includes a dedicated classifier and safety system prompt designed to reject requests to share highly sensitive information, such as credit card numbers.
Anthropic describes Claude Duet as an early version that may misread context or misjudge urgency. It notes limitations in handling personal errands like booking tickets. Users interested in trying the agent can join the waitlist.
- Oct 21, 2026 · Latent Space — swyx
Claude Mods introduce programmable human-facing agent interfaces
Trust78 - Oct 21, 2026 · Latent Space — swyx
DeepSeek open-sources DeepInfer inference engine with production-scale throughput and cost claims
Trust74 - Oct 21, 2026 · Latent Space — swyx
Tenex details agent-assisted SDLC workflow and compensation model for enterprise AI teams
Trust79