Coverage · growing
A standard that is filling in.
The tone corpus is not a frozen file — it grows as native contributors record. Here is the current, live curated footprint, and why its completeness is measurable.
6Languages
146Homographs
252Distinct meanings
327Consented takes
327Extractable clips
459FSIs assembled
| Language | Country | Homographs | Meanings | Takes | Clips | FSI size |
|---|---|---|---|---|---|---|
| 🇿🇲 Bemba | Zambia | 52 | 60 | 60 | 60 | 520 |
| 🇸🇿 Swati | Eswatini | 20 | 46 | 60 | 60 | 405 |
| 🇿🇦 Xhosa | South Africa | 25 | 44 | 60 | 60 | 455 |
| 🇿🇦 Zulu | South Africa | 29 | 58 | 60 | 60 | 460 |
| Swahili | Rwanda | 10 | 24 | 48 | 48 | 1000 |
| 🇷🇼 Kinyarwanda | Rwanda | 10 | 20 | 39 | 39 | 490 |
How to read this
Bounded denominator, rising numerator.
Because the syllable inventory of each Bantu language is finite and closed, coverage is a measurable property: a known set of legal syllables and homographs, filling with consented native audio. That is the opposite of an open-ended web scrape — you can see exactly how complete it is, and watch the number rise.