GitHub makes Copilot Memory generally available and previews opt-in dreaming
GitHub introduced its intelligence context and learning layer on October 28, combining memory, background context updates and code search across repositories. Availability and settings differ across those capabilities; the announcement does not make every component generally available.
Copilot Memory is generally available on paid Copilot plans. It can carry personal preferences and repository conventions across sessions and clients. GitHub lists the Copilot app, CLI, VS Code, cloud agent, code review, Slack, Teams and GitHub Mobile among places where memories can be created.
Users can interact with memories in the app, CLI and VS Code, and can view or delete personal memories in Copilot settings. Memories can capture details such as a preferred pull-request description format or a repository test command, reducing the need to repeat that guidance.
Copilot dreaming is a separate public preview. It can curate personal memories, update repository memories and turn recurring accepted code-review feedback into guidance. Changes to checked-in skills and instruction files are proposed through pull requests for team review.
Dreaming is opt-in, with personal and repository settings for individual subscribers and administrative policies for organizations and enterprises. Repository access is constrained by Copilot Memory settings, and users must also enable personal dreaming themselves. Dream runs consume AI credits.
Code search can find text, symbols and semantic matches across related repositories without a local clone, including through the /research command. GitHub also reports better preference adherence in 93% and more efficient workflows in 42% of reviewed preview CLI sessions where memory was used. These are vendor-reported observations from that subset, not universal performance gains or independent guarantees.