We record spontaneous speech from elders in Izzi, Ezaa, Ikwo, and Mgbo: life stories, market talk, proverbs, farming knowledge. Izzi alone has roughly 540,000 speakers and zero datasets anywhere. The method follows Kallaama, a community team that collected 125 hours under a Lacuna grant with nothing more than phones, consent forms, and local speakers.

Two funding doors are open now. ELDP small grants offer up to EUR 10,000 per project, with the next deadline on 1 October 2026. The Lacuna Fund NLP rounds are recurring and open to African-headquartered teams. A pilot with the ELDP becomes the evidence package for Lacuna.

Decades of professionally edited text in Igbo, Edo, Efik, Ibibio, Yoruba, and Hausa sit inside past exam papers and the WAEC selected WASSCE texts list. Nobody has formally asked for a research license to digitize this material. We are writing that letter. The ask is narrow: a non-commercial research license, anonymized past papers, attribution to WAEC, and co-branding on the resulting open dataset. Cost to try is one letter and one meeting.

Once the elder-voice pilot produces real assets, we build an API serving the Ebonyi cluster, modeled on the Igbo API rather than competing with it. Igbo API already accepts dialect contributions and tracks Izzi, Ezaa, and Edda as pronunciation variants. A partnership conversation is the natural first step. Engineering work without data behind it serves no one, so this follows the pilot.

Nigeria's film industry produces roughly 2,500 films a year (WIPO), many with scripts and subtitles in Nigerian languages. No open corpus of this material exists. We start the relationship with producers and guilds now. Licensed access to that scale of natural dialogue is a year-two or year-three result at the earliest, and the conversation is worth starting early.