Open data: recordings

• JSON: https://fuzhounese.org/api/export/recordings
• CSV, for a spreadsheet: https://fuzhounese.org/api/export/recordings?format=csv

Each response is the complete current list of published recordings, refreshed at most once an hour. A recording appears once an editor has listened to it.

Once a day is plenty. Each time:

• download the audio for every id you do not have yet, from its audio_url;
• update the details of the ones you have (a note or a credit can change);
• remove every recording that is no longer in the list. Takes are sometimes withdrawn, and contributors may ask to have their voice taken down; the list is the record of what may still be used.

The audio files are WebM (Opus), as phones and browsers record them; convert them if your player needs MP3 or AAC. The files are stored in the United States and can be slow to reach from mainland China, so keep your own copies rather than linking to ours.

• id, kind (the word itself, or an example sentence), audio_url, format, seconds
• word: the characters, the romanization as the dictionary writes it (Bàng-uâ-cê), the same in Yngping 榕拼 where it converts, the first English meaning and Mandarin gloss, and a link to the word's page
• sentence: for an example recording, the sentence read and its translation
• speaker, recorded_by, origin: who is speaking (when it is not the account holder), the account that recorded it, and where the speaker's Fuzhounese is from
• attribution: a ready-made credit line
• published_at, checked_at

The recordings are licensed CC BY-SA 4.0. Wherever a recording is played or published, credit the speaker and fuzhounese.org — the attribution field is written for that — and share anything you make from them under the same licence. Each contributor agreed to this licence when they recorded (terms).

Using the feed, or need something it does not have? Write to noahdlee519@gmail.com.