Anthropic announced the Claude Code Projects experience to manage long-term software development work. The system brings together multiple AI agents around a shared goal, memory and files, while Claude directs tasks across different work sessions and coordinates the results. Sessions can carry out work such as analysis, optimization and testing simultaneously.
Shared memory and a project library across sessions ensure that decisions and files are retained. Each session works on a separate copy and branch of the code repository; conflicts need to be resolved through standard merging processes. Projects uses the existing subscription quota and is initially being opened in beta to selected Pro and Max subscribers; access is planned to be expanded later.
Why it matters
This approach moves AI in software development beyond being a tool that responds to individual tasks and places it within a project workflow with continuity. Shared memory and the project library reduce the need to reconstruct decisions made in different sessions, while making it possible to divide work into specialized components such as analysis, optimization and testing. In return, teams will need to manage the burden of merging results produced in separate code copies and branches; therefore, efficiency depends not only on agents working simultaneously but also on how smoothly those results are merged. The fact that initial access is limited to selected Pro and Max subscribers means that the impact of the implementation will, for now, be assessed through the experiences of this user group.
Background
Anthropic is not a new name in the FikirPilot archive: we have published 27 news articles mentioning this name in the last 90 days; the most recent is dated September 17, 2026.