Hometecnologia · Abomey-Calavi, Bénin

Democratizing AI
in local languages.

HMTC brings artificial intelligence closer to everyday people by making it work in the languages they speak — starting with Maïka-AI for African languages.

Use Maïka-AI → Explore the API

12African languages

740M+potential speakers

v1language integration

scroll ↓

Coverage

Local languages are the interface.

Tap a colored country to see which Maïka-AI languages are spoken there and the potential user reach for each language.

Maïka-AI language footprint Other African countries

0 languages

0 potential speakers

0 countries

Product

Meet Maïka-AI — our evolving AI.

Maïka-AI is an evolving artificial intelligence built to operate in local languages. Its current stage focuses on integrating African languages into practical NLP and speech workflows, beginning with number and money-amount understanding.

19languages supported by v1 logic

15languages with live audio

645audio atoms (open)

Version 1 use cases

PaymentsRead transaction amounts aloud in the customer's local language at POS and merchant terminals.

Fintech alertsGenerate spoken confirmations for mobile money, agent banking, savings groups, and receipts.

Voice-first appsGive health, agriculture, education, and public-service tools a local-language speech layer.

Developer APIConvert amounts into language-specific words and audio atoms through a server-side API.

Use Maïka-AI → Read the docs

Developers

Build with Maïka-AI v1.

One public endpoint to compose money amounts into African-language text and audio atoms.

POST /api/compose

# Example API call
curl -X POST https://maika.hmtc.bj/api/compose \
  -H "Content-Type: application/json" \
  -d '{"lang":"fon","amount":7500}'

Amount-to-WordsCompose amounts with language-specific numeration rules.
Audio atomsReturn ordered word pieces that map to recorded WAV files.
Payment speechMake transaction confirmations understandable in local languages.
Conversational AIBroader local-language assistants. Evolution roadmap.

Get API access →

Open science