In its own words: Chikete
Need Kete data for AI? Speech for ASR and TTS, text and parallel data for LLMs, numbers, grammar, benchmarks — BantuNomics curates Bantu languages systematically, from consented native speakers. Tell us what you need: a full subscription lets us prioritise your languages of interest too.
Request Kete data Plans The method, live: numbers to a trillion in 20 languages · L26 benchmarkOne stem, kete — the beginning changes and the meaning moves from the language to a person, the people and the land.
From the BantuNomics language registry. Speak Kete and see something to correct? Tell us →
Everything here is built with people who speak the language, with consent. Speak Kete? Contribute at amina.ai.
Kete's complete operating alphabet is released and licensed from the Validation Pilot.
ReleasedThe class system above, the verb and the agreement it drives — licensed per language.
See plansRule-driven generation for Kete is not public yet — request it below.
Not yetWe curate Bantu languages on our own roadmap; requests tell us where labs need us. A Full Annual Subscription lets us prioritise your languages of interest too.
POST https://bantunomics.com/api/v1/language-requests
{"iso": "kcv", "domains": ["nouns","audio"],
"agent_name": "…", "contact_email": "…", "use_case": "…"}
Can only fetch (GET)? Fill the {…} and fetch:
https://bantunomics.com/api/v1/language-requests/new?iso=kcv&domains={domains}&agent_name={agent_name}&use_case={use_case}&confirm=yes
Or connect the MCP server https://bantunomics.com/languages/mcp and call request_language. Machine-readable page: /languages/kete.json