GitHub intelligence puts your code, context, and history to work
According to GitHub. GitHub announces a new intelligence layer designed to bring code context, repository conventions, and developer history into agent workflows.
- • GitHub intelligence serves as a context and learning layer sourced from GitHub data.
- • Copilot Memory is now generally available on paid Copilot plans across various apps and tools.
- • Copilot dreaming is introduced in public preview to continuously improve agentic sessions.
- • GitHub code search enables Copilot to find relevant code and documentation without a local clone.
Context and learning for developer agents
GitHub intelligence functions as a context and learning layer sourced from the graph of GitHub data. By leveraging team preferences, repository conventions, and development history, the system aims to help agents remember past work and adhere to developer standards across sessions.
Memory and dreaming capabilities
Copilot Memory is now generally available on paid Copilot plans, enabling agents to remember personal preferences and repository conventions. Additionally, Copilot dreaming is available in public preview to continuously reflect on past work and curate personal or repository memories.
Searching beyond the local checkout
GitHub code search allows Copilot to reach beyond a local checkout to find relevant code, configurations, schemas, and documentation across related projects and dependencies. It supports exact text matching, regular expressions, and semantic search based on what code actually does.
- 1. github.blogOpen source