Everything here can be read, checked, and argued with.

The code

The whole app lives in one repository: github.com/kihea/superb — the reading app, the learning engine (Rust, compiled to WebAssembly), this website, and the pipelines that build the content. It is licensed MIT with the Commons Clause: read it, run it, change it, share changes, but don't sell it as a service.

The books

All 614 books are public domain, kept in their own repository at superb-catalogue/library. Every book carries a provenance record — where the text came from, which edition, and the licence basis — complete enough that a stranger can verify it without asking anyone.

The word data

Word meanings come from the English Wiktionary (via wiktextract, CC BY-SA 4.0), rhyme judging from CMUdict, and word associations from WordNet and the project's own corpus. Each source is named with its licence in the repository's data manifest, and nothing with a non-commercial or all-rights-reserved licence enters a build.

How changes happen

Through ordinary GitHub issues and pull requests. If a definition is wrong, a passage reads badly, or a screen misbehaves, open an issue — plain words are enough, no template required. The repository's CONTRIBUTING file covers running it locally.

What transparency means here

Not everything is explained on every screen — the app aims for a minimal feel — but nothing is hidden. Every claim the product makes (what is fetched, what is stored, what the engine learns from) is checkable in the source, and the site's own build proves its network policy on every push.