Data Terms
Terms for the midicn-lib dataset — no rights claimed, per-track source licences, non-commercial
1. What this site is
These sites (the lib / mid / zip subdomains) are the browsing and download front-end for the midicn-lib dataset.
midicn-lib is not an original collection — it is an aggregator. We collect MIDI files published elsewhere under open licences (public domain, Creative Commons and similar), clean them, de-duplicate them, and organise them into a structured dataset for searching and research.
This matters: everything below rests on the fact that these works are not ours.
2. We claim no rights
We claim no rights whatsoever over the works in this dataset.
- We are not the copyright holders
- We cannot grant a blanket licence for them
- We do not alter their original licences
- We charge nothing
The terms of use for any given piece are its own source licence — every page states the source and licence marker.
3. How you may use it
Because the dataset is a collection of openly licensed content, most works can be used freely — but the exact conditions differ per work:
| The work is marked | What you must do |
|---|---|
| Public domain / CC0 | Essentially unrestricted |
| CC BY / CC BY-SA | Give attribution (BY-SA also requires share-alike) |
| CC BY-NC / NC-SA | Non-commercial use only |
| Traditional-music transcription (study tier) | Study and research only |
How to check: look at the licence marker on the track page, or consult the dataset's licence ledger.
We add no restrictions of our own — we are in no position to, and there is no need.
4. Our commitment: non-commercial
These sites carry no advertising, no paywalls and no commercial activity of any kind.
The reason is simple: what we publish here is other people's work. Monetising it would make the whole chain of compliance fragile. We do not do that.
This differs from midicn.com and the knowledge sites: those publish content we wrote ourselves (under CC BY 4.0), so they may be commercialised. The two groups follow different rules — read each site accordingly.
5. Please don't
- ❌ Scrape at high frequency — the dataset can be downloaded in full (see the downloads page); scraping only strains the hosting platform
- ❌ Strip or alter licence and source information — that breaks the upstream authors' terms and may infringe
- ❌ Claim the dataset as your own original work — use it freely, but don't misrepresent where it came from
- ❌ Use it for anything unlawful
6. Our compliance work
To keep the "carrying" itself legitimate, we do the following:
- Per-source licence checks — every source is verified before inclusion
- Removal logs — works from one source that appeared to be modern in-copyright material have been removed and documented (see the data copyright page)
- Per-track labelling — every entry carries its source and licence
- Traceability — the dataset's sources and changes are recorded
7. Changes and contact
These terms may be updated as the dataset evolves; a revised version replaces this page and is dated.
For licensing questions, mislabelled entries, or rights claims: open an issue on our GitHub repository.
This page is a plain-language summary, not legal advice.
Sources
- 本站自订条款，据实描述数据集的实际授权状况
Used for fact-checking only; all prose is original.