Anthropic Upgrades Claude Code Projects to Run Parallel AI Workers Hands-Free

Why it matters

Anthropic has upgraded Claude Code Projects, its system for running many cloud-based coding sessions in parallel, with features that reduce how closely developers must supervise those sessions. A coordinator, which assigns work to independent threads, now posts status updates and answers questions in project chat without being prompted, surfacing blocked work and pending decisions that previously required scanning a thread list. Individual threads can be shared as read-only links with signed-in Claude users, though threads that touched a local computer stay unshareable to avoid exposing machine data. Reminders let a thread hide itself and reappear at a scheduled time, while new slash commands support side conversations and branching. The threads panel adds a recents row and artifact thumbnails for easier retrieval. The public beta remains limited to Pro and Max subscribers, with Team and Enterprise excluded, and parallel workers draw from existing plan allowances, so heavy parallel use can exhaust limits faster than a single session. Anthropic positions Projects for divisible engineering work reconciled through branches and pull requests.