Changelog automation
Release notes assembled from pull requests, labels, or commit conventions, so each release starts with a useful draft instead of a blank page.
What it is
Changelog automation turns structured development history into release notes. A tool reads pull request titles, labels, or Conventional Commits, groups the changes, and produces the draft people edit or publish. GitHub can generate release notes from merged pull requests, while Changesets can also update package versions and CHANGELOG files in a monorepo. The Added, Changed, and Fixed headings in so many CHANGELOG.md files come from Keep a Changelog, the format most of these tools emit.
Reach for it when releases are frequent enough that hand-written notes lag behind, or when nobody can reconstruct a month of changes on release day. The useful unit is usually the pull request because it can carry a human title, category label, author, and link.
Gotcha: automation preserves bad inputs with great consistency. Titles such as 'fix stuff' and missing breaking-change labels produce useless notes. Require a reader-facing pull request title and one release category, exclude chores and dependency noise deliberately, and keep a human review before publication.
Ask AI for it
Set up GitHub generated release notes for this repository. Add .github/release.yml with categories for breaking changes, features, fixes, and documentation based on exact pull request labels, plus an exclude section for skip-changelog and dependency-only updates. Add a pull request template that requires a reader-facing summary and a release-note label, then add a GitHub Actions check that fails when neither a category label nor skip-changelog is present. Document how to create a GitHub Release with Generate release notes and require human review before publishing.