Available Now
Bantu Embeddings API
Vector embeddings purpose-built for Africa's languages
Generate high-quality vector embeddings for Bantu text. Power search, classification, and RAG in 22 African languages with a single API call.
Features
What you get
256 and 512-dimensional vectors
Contrastive learning on Bantu language data
Sub-100ms latency
Supports IsiZulu, IsiXhosa, KiSwahili, ChiShona, and more
Powered by proprietary Bantu language models
REST API — simple POST /embed
Pricing: $0.02 per 1M tokens
Request
curl -X POST https://api.bhala.ai/v1/embed \
-H "Authorization: Bearer bh_sk_..." \
-H "Content-Type: application/json" \
-d '{
"input": "Sawubona, ngingakusiza kanjani?",
"model": "bantu-embed-v1",
"dimensions": 256
}'Response
{
"object": "embedding",
"data": [
{
"index": 0,
"embedding": [0.0234, -0.0891, 0.1456, ...],
"dimensions": 256
}
],
"model": "bantu-embed-v1",
"usage": {
"prompt_tokens": 8,
"total_tokens": 8
}
}Use Cases
What you can build with Bantu Embeddings API
Semantic search across multilingual documents
Retrieval-Augmented Generation (RAG) in Bantu languages
Text classification and clustering
Duplicate detection and content matching
Language Support
Supported languages
Purpose-built for the Bantu language family with more languages being added continuously.
IsiZulu
12M+ speakers
IsiXhosa
8M+ speakers
Sepedi
5M+ speakers
Setswana
5M+ speakers
Sesotho
6M+ speakers
Xitsonga
4M+ speakers
IsiSwati
2.5M+ speakers
Tshivenda
1.3M+ speakers
Lingala
20M+ speakers
Tshiluba
6M+ speakers
Kikongo
5M+ speakers
KiSwahili
15M+ L1 speakers
Kinyarwanda
13M+ speakers
ChiShona
11M+ speakers
IsiNdebele
2M+ speakers
ChiTonga
1.5M+ speakers
Umbundu
6M+ speakers
Kimbundu
3M+ speakers
Kirundi
11M+ speakers
Chichewa
10M+ speakers
Bemba
4M+ speakers
Luganda
5.5M+ speakers