What happened
On March 23, 2026, OpenAI rolled out the Library feature to ChatGPT for Plus, Pro, and Business users globally (with EEA, Switzerland, and UK availability pending). Files uploaded or created in ChatGPT — PDFs, spreadsheets, images, generated outputs — are now automatically saved to a persistent Library rather than disappearing at the end of a session. The Library is accessible via a sidebar tab and supports natural language search across saved files. Recent files appear in the composer for quick reuse, and generated images are organized in a dedicated Images tab. Files uploaded in Temporary Chat mode are not saved.
Why it matters
Previously, every ChatGPT conversation started with a blank slate — documents uploaded in one session were unavailable in the next, forcing users to re-upload context repeatedly. Library changes the interaction model toward a persistent workspace where prior context accumulates. For knowledge workers who use ChatGPT regularly with recurring reference documents, this meaningfully reduces friction. The natural language search over saved files hints at a longer-term direction: ChatGPT as a personal knowledge base, not just a turn-by-turn chat interface. This is an incremental but meaningful step toward stateful AI work environments.
Who should pay attention
- Business users and knowledge workers who frequently reference the same documents across multiple ChatGPT sessions
- Teams evaluating ChatGPT for workflow integration where persistent context matters
- Developers building on the ChatGPT API who may see analogous stateful features appear in API offerings