Copilot Spark giving enterprise users a vibe coding editor inside GitHub is the feature most developers have not heard about

A vibe coding editor that generates programs from text prompts, available in public preview for GitHub Enterprise, is different from Copilot's existing code completion and chat features. It is a generation-from-description tool built into the GitHub environment rather than into an IDE.

The Copilot agent using the entire conversation history within a pull request for improved context is the quality improvement that addresses the most common frustration with AI code review: responses that do not account for the decisions already discussed in the PR thread.

The ability to prompt directly on GitHub.com notifications for task creation and code changes is the ambient accessibility that reduces the tool-switching overhead of managing development work across platforms.

For GitHub Enterprise users specifically: has anyone tried Copilot Spark for generating programs from text prompts and how does the output quality compare to using Copilot in an IDE?