Get involved
Whoever you are, there is a way in
Nigeria has over 500 languages. Research covers fewer than 1%. The Ebonyi cluster sits deep in that gap. Getting out of it takes speakers, institutions, money, code, and people who understand the communities involved. Every role below is real and currently open.
Speakers and elders in Ebonyi
Izzi alone has around 540,000 speakers and zero published datasets. Your voice is the data. One session recording life stories, farming, proverbs, or a church reading produces material no server anywhere holds today.
The New Testaments in Izii, Ezaa, and Ikwo were finished in 1980. The world wrote these languages down. Now we need machines to hear them. Recording sessions run at palace grounds and village halls, with oral and written consent in your language before anything is captured.
Only 2.4% of new African voice data comes from speakers aged 50 and above. Elders are the gap. Showing up is the fix.
Churches, town unions, traditional rulers
Faith Comes By Hearing and Global Recordings Network already produced free audio scripture in Izii, Ezza, Ikwo, and Ngbo. This work continues that. Sermons, testimonies, and interpreted readings recorded with a pulpit announcement and a consent corner after service give us hours of natural continuous speech.
Traditional rulers and town unions hold something no grant can buy: the trust of speakers. A chief opening a recording day, or a union secretary sending the WhatsApp message, brings in families who would not otherwise come. The linguistic survey of Ebonyi State confirms these languages are distinct from Igbo. Cultural custodianship over that distinctness is yours.
We ask for three things: an announcement from the pulpit or the palace, access to a hall for one day, and permission to store recordings in a permanent archive at ELAR under your community's agreed terms.
Diaspora and town unions abroad
Igbo API, YorubaName, and NaijaVoices were all built with diaspora energy and money. A town union abroad can fund one village recording day for roughly what a table at a gala costs, and the output outlasts the event by decades.
If you have elders at home who will not travel, a WhatsApp video call with a family member holding a phone is a valid recording session when done with proper consent. The Aikuma methodology was designed for exactly this: remote oral documentation with no specialist equipment.
Diaspora donations also unlock grant matches. ELDP small grants of up to EUR 10,000 are open to groups of any nationality with an October 1 deadline. A letter of community backing from a town union strengthens an application more than any academic credential.
Developers and researchers
The Igbo API dialect contribution docs already accept Izzi and Ezaa entries. One developer built the first Igbo dictionary API and proved that a small team can move a language from invisible to searchable. We are looking for people who want to do the same for the Ebonyi cluster.
On the research side, Masakhane has built the largest African NLP community working in the open. Their Slack is the place to find collaborators, get methodology feedback, and submit to the AfricaNLP workshop. Fine-tuning Whisper on NaijaVoices data cut word-error rate by roughly 76%. The Ebonyi cluster has no baseline yet, which means the first paper here is a genuine contribution to the record.
We need people who can run transcription sprints, structure metadata schemas, and eventually build the API layer that serves our recorded data. If that is you, write to us with what you can contribute and what you need from the project in return.
Health, agriculture, and government partners
Speech-first interfaces are not an experiment. Meridian Hospital in Enugu cut radiology turnaround from 48 hours to 20 minutes after deploying a voice documentation system. Intron Health builds these systems for African clinical settings and raised $1.6M in 2024 to scale them. None of it works in Izzi or Ezaa today. The data we collect is the prerequisite.
On the government side, India's Bhashini platform shows what a national language-data infrastructure looks like when government is a genuine partner: prompt-text supply, radio archive access, and small matching grants that turn community projects into national assets. Rwanda did the same for Kinyarwanda and it became the number two voice dataset on Mozilla Common Voice. Ebonyi's ministries of Culture, Education, and ICT are the natural counterparts here.
Agriculture extension workers, health educators, and ward development committees already communicate in these languages. Recording those interactions, with consent, is a data collection method and a service improvement at the same time. If you run a program that touches Ebonyi communities, let us talk about what a partnership looks like.
On the ground
What our week looks like when you help
This is not a hypothetical. It is the rhythm we run now, with the volunteers and partners already on board. More people in any of the five roles above make every one of these items faster and deeper.
Monday
Proverb post
One Izzi or Ezaa proverb goes out with audio and translation. We recorded it the week before. Takes thirty minutes of editing.
Tuesday to Thursday
Transcription sprint
Student volunteers from EBSU or AE-FUNAI Ikwo work through the previous weekend's recordings. Each session yields three to five hours of clean, timestamped text.
Friday
Metadata check
Speaker age, village, language variety, and consent status are verified for every file before it enters the archive. Nothing ships without a complete record.
Saturday
Field day (when scheduled)
A palace or hall, a chief opening the session, twenty to fifty speakers at stations: life stories, farming talk, proverbs, free conversation. One day yields ten to thirty recorded hours.
Sunday
Partner call
A short call with a church contact, a diaspora union secretary, or a researcher. Most of the project's relationships are built one call at a time.
Want to see the full collection methodology? Read the Kallaama paper (125 hours of Wolof speech collected by a team of similar size) and the Aikuma documentation tools. That is the playbook we follow.
“Onye wetara oji wetara ndu”
The one who brings kola brings life