FSI integration

The syllable is the tone-bearing unit.

Tone docks on exactly one place — the vowel of each syllable. So you cannot represent tone without first decomposing the word into its syllables, and you cannot do that without the language's Full Syllable Inventory. BantuNomics has assembled 459 of them — the first complete collection in history.

Read why FSIs Hear the data

🇿🇲 Bemba 🇸🇿 Swati 🇿🇦 Xhosa Swahili 🇷🇼 Kinyarwanda

Minimal pairs

Same syllables. Different tone. Different meaning.

Each card is one written word that the FSI decomposes into the same syllable sequence for every meaning. The spelling is identical; the tone — carried in the audio, on the syllable nuclei — is what separates the meanings. Play them back to back.

Bwana 3 meanings

bwa·na

Husband

Bwana

means husband

Audio pending.

Native pronunciation note: Bwah-nah

{
 "take_id": "4e627f0c-dd82-4d68-b013-95ec194811bd",
 "language_iso": "swh",
 "kind": "word",
 "bantu": "Bwana",
 "english_gloss": "Husband",
 "switch_ms": 4667,
 "duration_ms": 9870,
 "segments": [
  {
   "lang": "swh",
   "t0_ms": 0,
   "t1_ms": 4667
  },
  {
   "lang": "en",
   "t0_ms": 4667,
   "t1_ms": 9870
  }
 ],
 "speaker": "Speaker SWH-A",
 "consent": "granted"
}

Master or Lord

Bwana

means master or lord

Audio pending.

Native pronunciation note: Bwah-nah

{
 "take_id": "eff6555e-8c53-4669-9e08-63ac0af00ff7",
 "language_iso": "swh",
 "kind": "word",
 "bantu": "Bwana",
 "english_gloss": "Master or Lord",
 "switch_ms": 4669,
 "duration_ms": 10800,
 "segments": [
  {
   "lang": "swh",
   "t0_ms": 0,
   "t1_ms": 4669
  },
  {
   "lang": "en",
   "t0_ms": 4669,
   "t1_ms": 10800
  }
 ],
 "speaker": "Speaker SWH-A",
 "consent": "granted"
}

Mister or Sir

Bwana

means mister or sir

Audio pending.

Native pronunciation note: Bwah-nah

{
 "take_id": "a0ddd8e1-cf85-421f-94d0-618a6ee07757",
 "language_iso": "swh",
 "kind": "word",
 "bantu": "Bwana",
 "english_gloss": "Mister or Sir",
 "switch_ms": 4666,
 "duration_ms": 11460,
 "segments": [
  {
   "lang": "swh",
   "t0_ms": 0,
   "t1_ms": 4666
  },
  {
   "lang": "en",
   "t0_ms": 4666,
   "t1_ms": 11460
  }
 ],
 "speaker": "Speaker SWH-A",
 "consent": "granted"
}

Karibu 3 meanings

ka·ri·bu

Almost

Karibu

means almost

Audio pending.

Native pronunciation note: Kah-ree-boo

{
 "take_id": "8fa1ae78-ae02-409f-ba3d-b3647146002a",
 "language_iso": "swh",
 "kind": "word",
 "bantu": "Karibu",
 "english_gloss": "Almost",
 "switch_ms": 5148,
 "duration_ms": 8850,
 "segments": [
  {
   "lang": "swh",
   "t0_ms": 0,
   "t1_ms": 5148
  },
  {
   "lang": "en",
   "t0_ms": 5148,
   "t1_ms": 8850
  }
 ],
 "speaker": "Speaker SWH-A",
 "consent": "granted"
}

Near or close to

Karibu

means near or close to

Audio pending.

Native pronunciation note: Kah-ree-boo

{
 "take_id": "a17f0368-c451-4f2b-89ed-14fa67d3d6cf",
 "language_iso": "swh",
 "kind": "word",
 "bantu": "Karibu",
 "english_gloss": "Near or close to",
 "switch_ms": 5148,
 "duration_ms": 10820,
 "segments": [
  {
   "lang": "swh",
   "t0_ms": 0,
   "t1_ms": 5148
  },
  {
   "lang": "en",
   "t0_ms": 5148,
   "t1_ms": 10820
  }
 ],
 "speaker": "Speaker SWH-A",
 "consent": "granted"
}

Welcome

Karibu

means welcome

Audio pending.

Native pronunciation note: Kah-ree-boo

{
 "take_id": "c6e59b32-1188-4dae-8c26-1d51057f6a4c",
 "language_iso": "swh",
 "kind": "word",
 "bantu": "Karibu",
 "english_gloss": "Welcome",
 "switch_ms": 5152,
 "duration_ms": 9390,
 "segments": [
  {
   "lang": "swh",
   "t0_ms": 0,
   "t1_ms": 5152
  },
  {
   "lang": "en",
   "t0_ms": 5152,
   "t1_ms": 9390
  }
 ],
 "speaker": "Speaker SWH-A",
 "consent": "granted"
}

Paka 3 meanings

pa·ka

Cat

Paka

means cat

Audio pending.

Native pronunciation note: Pah-kah (stress on the first syllable)

{
 "take_id": "59ac1d71-0ab0-47a4-9015-2c8d9b78a43d",
 "language_iso": "swh",
 "kind": "word",
 "bantu": "Paka",
 "english_gloss": "Cat",
 "switch_ms": 4669,
 "duration_ms": 9180,
 "segments": [
  {
   "lang": "swh",
   "t0_ms": 0,
   "t1_ms": 4669
  },
  {
   "lang": "en",
   "t0_ms": 4669,
   "t1_ms": 9180
  }
 ],
 "speaker": "Speaker SWH-A",
 "consent": "granted"
}

To paint

Paka

means to paint

Audio pending.

Native pronunciation note: Pah-kah

{
 "take_id": "46c941aa-721f-4ee4-a1d5-e52a1850467f",
 "language_iso": "swh",
 "kind": "word",
 "bantu": "Paka",
 "english_gloss": "To paint",
 "switch_ms": 4669,
 "duration_ms": 12930,
 "segments": [
  {
   "lang": "swh",
   "t0_ms": 0,
   "t1_ms": 4669
  },
  {
   "lang": "en",
   "t0_ms": 4669,
   "t1_ms": 12930
  }
 ],
 "speaker": "Speaker SWH-A",
 "consent": "granted"
}

To smear

Paka

means to smear

Audio pending.

Native pronunciation note: Pah-kah

{
 "take_id": "b328ec87-9307-4b23-beb8-63815009a850",
 "language_iso": "swh",
 "kind": "word",
 "bantu": "Paka",
 "english_gloss": "To smear",
 "switch_ms": 4670,
 "duration_ms": 11140,
 "segments": [
  {
   "lang": "swh",
   "t0_ms": 0,
   "t1_ms": 4670
  },
  {
   "lang": "en",
   "t0_ms": 4670,
   "t1_ms": 11140
  }
 ],
 "speaker": "Speaker SWH-A",
 "consent": "granted"
}

Panda 3 meanings

pa·nda

To climb

Panda

means to climb

Audio pending.

Native pronunciation note: Pahn-dah

{
 "take_id": "a4e50266-7db3-4669-9197-9ec1d98001b0",
 "language_iso": "swh",
 "kind": "word",
 "bantu": "Panda",
 "english_gloss": "To climb",
 "switch_ms": 4664,
 "duration_ms": 8860,
 "segments": [
  {
   "lang": "swh",
   "t0_ms": 0,
   "t1_ms": 4664
  },
  {
   "lang": "en",
   "t0_ms": 4664,
   "t1_ms": 8860
  }
 ],
 "speaker": "Speaker SWH-A",
 "consent": "granted"
}

To go up or ascend

Panda

means to go up or ascend

Audio pending.

Native pronunciation note: Pahn-dah

{
 "take_id": "e33ee4e4-cc91-463b-91fb-92a359eed84f",
 "language_iso": "swh",
 "kind": "word",
 "bantu": "Panda",
 "english_gloss": "To go up or ascend",
 "switch_ms": 4670,
 "duration_ms": 10570,
 "segments": [
  {
   "lang": "swh",
   "t0_ms": 0,
   "t1_ms": 4670
  },
  {
   "lang": "en",
   "t0_ms": 4670,
   "t1_ms": 10570
  }
 ],
 "speaker": "Speaker SWH-A",
 "consent": "granted"
}

To plant or sow

Panda

means to plant or sow

Audio pending.

Native pronunciation note: Pahn-dah

{
 "take_id": "08dd2813-ee56-473f-a513-b719296e3b3d",
 "language_iso": "swh",
 "kind": "word",
 "bantu": "Panda",
 "english_gloss": "To plant or sow",
 "switch_ms": 4670,
 "duration_ms": 10940,
 "segments": [
  {
   "lang": "swh",
   "t0_ms": 0,
   "t1_ms": 4670
  },
  {
   "lang": "en",
   "t0_ms": 4670,
   "t1_ms": 10940
  }
 ],
 "speaker": "Speaker SWH-A",
 "consent": "granted"
}

Baridi 2 meanings

ba·ri·di

Calm or peaceful

Baridi

means calm or peaceful

Audio pending.

Native pronunciation note: Bah-ree-dee

{
 "take_id": "90c47885-d0e5-49e6-a3bc-b80c7a63119b",
 "language_iso": "swh",
 "kind": "word",
 "bantu": "Baridi",
 "english_gloss": "Calm or peaceful",
 "switch_ms": 5148,
 "duration_ms": 12400,
 "segments": [
  {
   "lang": "swh",
   "t0_ms": 0,
   "t1_ms": 5148
  },
  {
   "lang": "en",
   "t0_ms": 5148,
   "t1_ms": 12400
  }
 ],
 "speaker": "Speaker SWH-A",
 "consent": "granted"
}

Cold

Baridi

means cold

Audio pending.

Native pronunciation note: Bah-ree-dee

{
 "take_id": "b20d0bcc-c161-48e0-a136-fc39168474e0",
 "language_iso": "swh",
 "kind": "word",
 "bantu": "Baridi",
 "english_gloss": "Cold",
 "switch_ms": 5146,
 "duration_ms": 9170,
 "segments": [
  {
   "lang": "swh",
   "t0_ms": 0,
   "t1_ms": 5146
  },
  {
   "lang": "en",
   "t0_ms": 5146,
   "t1_ms": 9170
  }
 ],
 "speaker": "Speaker SWH-A",
 "consent": "granted"
}

Hema 2 meanings

he·ma

Tent

Hema

means tent

Audio pending.

Native pronunciation note: Heh-mah

{
 "take_id": "5ac5991f-cfc9-495a-b790-3f5356e28d60",
 "language_iso": "swh",
 "kind": "word",
 "bantu": "Hema",
 "english_gloss": "Tent",
 "switch_ms": 4669,
 "duration_ms": 9050,
 "segments": [
  {
   "lang": "swh",
   "t0_ms": 0,
   "t1_ms": 4669
  },
  {
   "lang": "en",
   "t0_ms": 4669,
   "t1_ms": 9050
  }
 ],
 "speaker": "Speaker SWH-A",
 "consent": "granted"
}

To pant or breathe heavily

Hema

means to pant or breathe heavily

Audio pending.

Native pronunciation note: Heh-mah

{
 "take_id": "952883ab-eb8a-4ff9-ac85-d751d7412c46",
 "language_iso": "swh",
 "kind": "word",
 "bantu": "Hema",
 "english_gloss": "To pant or breathe heavily",
 "switch_ms": 4671,
 "duration_ms": 12430,
 "segments": [
  {
   "lang": "swh",
   "t0_ms": 0,
   "t1_ms": 4671
  },
  {
   "lang": "en",
   "t0_ms": 4671,
   "t1_ms": 12430
  }
 ],
 "speaker": "Speaker SWH-A",
 "consent": "granted"
}

Kaa 2 meanings

kaa

Coal

Kaa

means coal

Audio pending.

Native pronunciation note: Kaa-a (stretched or doubled a)

{
 "take_id": "14597b59-5fb8-4a4b-b310-5bff44ab27b6",
 "language_iso": "swh",
 "kind": "word",
 "bantu": "Kaa",
 "english_gloss": "Coal",
 "switch_ms": 4189,
 "duration_ms": 10450,
 "segments": [
  {
   "lang": "swh",
   "t0_ms": 0,
   "t1_ms": 4189
  },
  {
   "lang": "en",
   "t0_ms": 4189,
   "t1_ms": 10450
  }
 ],
 "speaker": "Speaker SWH-A",
 "consent": "granted"
}

To sit

Kaa

means to sit

Audio pending.

Native pronunciation note: Ka-a (stress on the first syllable)

{
 "take_id": "89f512cd-b994-423e-b3d9-b2474dfcc0a7",
 "language_iso": "swh",
 "kind": "word",
 "bantu": "Kaa",
 "english_gloss": "To sit",
 "switch_ms": 4190,
 "duration_ms": 8240,
 "segments": [
  {
   "lang": "swh",
   "t0_ms": 0,
   "t1_ms": 4190
  },
  {
   "lang": "en",
   "t0_ms": 4190,
   "t1_ms": 8240
  }
 ],
 "speaker": "Speaker SWH-A",
 "consent": "granted"
}

Kazi 2 meanings

ka·zi

Difficult or hard

Kazi

means difficult or hard

Audio pending.

Native pronunciation note: Kah-zee

{
 "take_id": "32660a64-5593-40b1-bdc9-f08c6f28543e",
 "language_iso": "swh",
 "kind": "word",
 "bantu": "Kazi",
 "english_gloss": "Difficult or hard",
 "switch_ms": 4670,
 "duration_ms": 10180,
 "segments": [
  {
   "lang": "swh",
   "t0_ms": 0,
   "t1_ms": 4670
  },
  {
   "lang": "en",
   "t0_ms": 4670,
   "t1_ms": 10180
  }
 ],
 "speaker": "Speaker SWH-A",
 "consent": "granted"
}

Work, job, or task

Kazi

means work, job, or task

Audio pending.

Native pronunciation note: Kah-zee

{
 "take_id": "6d4511e0-5ef6-474d-8593-ead9d632255f",
 "language_iso": "swh",
 "kind": "word",
 "bantu": "Kazi",
 "english_gloss": "Work, job, or task",
 "switch_ms": 4668,
 "duration_ms": 11820,
 "segments": [
  {
   "lang": "swh",
   "t0_ms": 0,
   "t1_ms": 4668
  },
  {
   "lang": "en",
   "t0_ms": 4668,
   "t1_ms": 11820
  }
 ],
 "speaker": "Speaker SWH-A",
 "consent": "granted"
}

Ndege 2 meanings

nde·ge

Airplane

Ndege

means airplane

Audio pending.

Native pronunciation note: Ndeh-geh

{
 "take_id": "312e85d0-3b9d-447c-b717-2493e428bb1f",
 "language_iso": "swh",
 "kind": "word",
 "bantu": "Ndege",
 "english_gloss": "Airplane",
 "switch_ms": 4669,
 "duration_ms": 9230,
 "segments": [
  {
   "lang": "swh",
   "t0_ms": 0,
   "t1_ms": 4669
  },
  {
   "lang": "en",
   "t0_ms": 4669,
   "t1_ms": 9230
  }
 ],
 "speaker": "Speaker SWH-A",
 "consent": "granted"
}

Bird

Ndege

means bird

Audio pending.

Native pronunciation note: Ndeh-geh

{
 "take_id": "cbf050a8-61b7-4ac3-a87d-9d9d7a2256c2",
 "language_iso": "swh",
 "kind": "word",
 "bantu": "Ndege",
 "english_gloss": "Bird",
 "switch_ms": 4671,
 "duration_ms": 8690,
 "segments": [
  {
   "lang": "swh",
   "t0_ms": 0,
   "t1_ms": 4671
  },
  {
   "lang": "en",
   "t0_ms": 4671,
   "t1_ms": 8690
  }
 ],
 "speaker": "Speaker SWH-A",
 "consent": "granted"
}

Tupa 2 meanings

tu·pa

A file, that is the abrasive metal tool used for smoothing

Tupa

means a file, that is the abrasive metal tool used for smoothing

Audio pending.

Native pronunciation note: Tu-pa (stress on the last syllable)

{
 "take_id": "35ae3125-63a0-4f86-ac0c-43dd1965da4c",
 "language_iso": "swh",
 "kind": "word",
 "bantu": "Tupa",
 "english_gloss": "A file, that is the abrasive metal tool used for smoothing",
 "switch_ms": 4670,
 "duration_ms": 14160,
 "segments": [
  {
   "lang": "swh",
   "t0_ms": 0,
   "t1_ms": 4670
  },
  {
   "lang": "en",
   "t0_ms": 4670,
   "t1_ms": 14160
  }
 ],
 "speaker": "Speaker SWH-A",
 "consent": "granted"
}

To throw away or discard

Tupa

means to throw away or discard

Audio pending.

Native pronunciation note: Tu-pa (stress on the first syllable)

{
 "take_id": "0d807c93-0030-4a53-aad1-468121c04b7b",
 "language_iso": "swh",
 "kind": "word",
 "bantu": "Tupa",
 "english_gloss": "To throw away or discard",
 "switch_ms": 4671,
 "duration_ms": 10740,
 "segments": [
  {
   "lang": "swh",
   "t0_ms": 0,
   "t1_ms": 4671
  },
  {
   "lang": "en",
   "t0_ms": 4671,
   "t1_ms": 10740
  }
 ],
 "speaker": "Speaker SWH-A",
 "consent": "granted"
}

Why a model can't fake this

Generation is cheap. Certification against native ground truth is not.

A model can output a plausible syllable list; it cannot prove the list is complete, correct, or native, and it cannot produce the consented audio that carries the tone. The FSI plus this audio is the certified standard underneath alignment, ASR, and TTS.