We’re launching Lyria 3.5 in Google Flow Music, with advances across musicality, lyrics, vocals, and creative control
deepmind.google·6d ago
AI Engineering
OpenAI Blog·Apr 10·5 min read
Using projects in ChatGPT
TL;DR
Users of ChatGPT previously lacked a way to organize multiple conversations, uploaded files, and custom instructions into coherent workspaces for ongoing projects. OpenAI built Projects, a feature that groups related chats, files, and instructions within a single container.
✦ Why It Matters
Engineers can now organize ChatGPT work by project, share context with collaborators, and maintain separate instruction sets per workflow.
Key Takeaways
- Group related conversations and files into named projects to reduce context switching overhead.
- Store custom instructions per project so AI behavior persists without manual reconfiguration.
- Share project access with team members to enable asynchronous collaboration on shared tasks.
- Use projects to isolate work by domain—code, documentation, research—for cleaner workflow separation.
Full Summary
ChatGPT users managing complex or recurring work faced friction: conversations scattered across the interface, files uploaded repeatedly, and custom instructions (system prompts that shape AI behavior) lost between sessions. OpenAI introduced Projects, a workspace feature that bundles chats, uploaded documents, and persistent instructions into named, reusable containers.
Each project maintains its own context and configuration, allowing users to switch between distinct work streams without losing state. The approach treats projects as first-class organizational units rather than flat conversation lists.
Engineers and teams can now isolate work by domain—e.g., separate projects for code review, documentation, or research—and share project access for collaborative workflows. This reduces setup overhead and improves continuity for long-running tasks.