The readings.
This page only shows what exists. Every number here is a file on our machines, shown as the thing it is. Where almost nothing existed for these languages, this is what exists now, and it is ours.
-
Paper · October 2026
Counting What Exists
Our first paper. A repeatable protocol for building language resources where almost none exist, with a first case from Nigeria: Pidgin, Izon and Urhobo. It counts what exists, credits it, and sets out the first paid round of speakers.
-
Yfraud · measured on
- 235,795
- web addresses
- 6,984
- phishing emails
- 10,000
- situations proved
- 50
- real cases labelled by hand
Scores and misses published, not summarised.
Measurement · public
Yfraud, measured
Rules decide. A model only explains. Every check runs on your own computer, and the measurement is on a public repository where the misses sit next to the scores.
-
Text · 19,000 headwords · credited to their authors
The Ijaw dictionaries
The Dictionary of Kolokuma Ịzọn by Kay Williamson and Roger Blench, and the Nembe dictionary, about 8,500 and 10,500 headwords, digitised and cleaned to text by us. Credit stays with their authors and the books stay theirs. The Ijaw lesson set is cut from that text, starting with nụ́à, the greeting on the first Creeks card.
-
Warri Pidgin wordlist · 336 entries · 293 glossed by a first language speaker
pompi calm down
pipe low maintain, same as reduce your voltage, bring your energy down
shege trouble, hardship, the bad side of things. You for see shege, you would have seen trouble
omo an exclamation in front of a sentence: wow, man, ah. Omo, na she come first
wetin sele what happened?
Growing
The Warri Pidgin wordlist
336 entries in the register people actually use in Warri, 293 of them glossed by a first language speaker in sessions where every meaning was given back in his own words. Kept separate from textbook Pidgin. The rows above are ones the open corpus never caught, and there is no other list like it.
-
tebraIjaw · how are you? · her voice, Warri · 0.65 s
The founding voice, recorded with consent and a share in the company, October 2026: 29 minutes of Ijaw and Pidgin, 73 items cut, 18 in the lesson today.
Round 1 · recording
South South voices
Round 1 is three speakers each in Pidgin, Urhobo and Ijaw, 6,000 items per language, recorded against a list of 200 words and 50 lines per clan. The first speaker is in and the list is written. Every count on this page is measured from the files, and every item carries the person who said it.
-
Open data · reference only
The Nigerian Pidgin treebank
The open Naija treebank of 9,241 sentences, used only as a reference for choosing words. Its sentences stay out of our data.
The research is the products.
We build things people use, and what those things learn becomes Y. The paper shows how we count and collect. The objects above grow because people use the products.
Get on the list.
Banks, fintechs and teams get Yfraud and Y first. Speakers and builders get Creeks. Leave your name and we will write to you.
You are on the list
Thank you, . We will write to when it is your turn.