Copilot code review now runs on an agentic architecture

Source: GitHub Changelog

What’s new? ✨

Copilot code review now uses agentic tool calling to gather broader repository context as needed (e.g., relevant code, directory structure, and references) so feedback reflects how your changes fit into the larger architecture. With agentic tool‑calling, you will now receive:

• Higher‑quality findings that prioritize correctness and architectural integrity.
• Lower noise, helping ensure comments surface meaningful signal rather than volume.
• More actionable guidance, helping developers resolve issues quickly and confidently.

Why it matters💡

You get more targeted, higher‑signal review comments with less noise. This helps you merge with confidence and ship cleaner code faster.

New to Copilot code review? Get started by checking out our Copilot code review documentation.

Read more about Copilot code review’s agentic architecture in this blog post.

Self-hosted runners: Setup required 🛠️

Agentic Copilot code review runs on GitHub Actions. If you have opted out of GitHub‑hosted runners, you’ll need to complete a one‑time self‑hosted runners setup to receive agentic Copilot code reviews on pull requests. Otherwise, there is no setup required. If you prefer, Copilot code review can also run on GitHub‑hosted larger runners.

Join the discussion within the GitHub Community.