ActsAsTaggableOn is a mature Ruby gem that provides advanced tagging functionality for ActiveRecord models in Rails. It allows models to be taggable on multiple named contexts (e.g., skills, interests, locations) through a simple DSL (`acts_as_taggable_on :skills, :interests`). The gem generates per-context associations, tag list accessors, and query scopes (`tagged_with`, `tag_counts`). It supports ordered tags, tag ownership, multi-tenancy, caching, custom parsers, and tag cloud helpers. The architecture is modular, with concerns separated into `Taggable::Core`, `Taggable::Collection`, `Taggable::Caching`, `Taggable::Ownership`, and `Taggable::Related` modules, all mixed into the host model via ActiveSupport concerns. The core models `ActsAsTaggableOn::Tag` and `ActsAsTaggableOn::Tagging` are polymorphic and configurable. Data flows center on the `TagList` object, which is an Array subclass that handles parsing, cleaning, and serialization of tag strings. Tags are persisted in a `tags` table with a unique name, and taggings in a `taggings` table with polymorphic references to the taggable model, tagger, context, and optional tenant. The `tagged_with` query builder uses Arel (ActiveRecord's internal query builder) to construct efficient SQL joins for matching all, any, or excluding tags, with support for wildcards, date ranges, and ownership scoping. External dependencies are minimal: ActiveRecord (>= 8.0, < 9.0) and zeitwerk for autoloading. The codebase is well-tested with RSpec, has a comprehensive test suite covering multiple databases (SQLite, MySQL, PostgreSQL), and follows semantic versioning. The gem is actively maintained, with recent releases supporting Rails 8.0/8.1 and Ruby 3.3+. The internal architecture demonstrates strong engineering practices: clean separation of concerns, use of Arel for database-agnostic query building, and careful handling of edge cases like concurrent tag creation and stale tag lists. Several components, such as the `TagList` class and the `DefaultParser`, are self-contained and could be extracted for use in other projects.