[ Company / Our Story ]

Why we exist. What we are building.

The journey of LocaleNLP from a small research project to a leading provider of African and Arabic language infrastructure — built by linguists and engineers from the communities we serve.

History

Our journey

Q1

Founded in Marrakech

~/history/2024
LocaleNLP founded by Alieu Jagne in Marrakech, Morocco. The vision: build the foundational language infrastructure layer for African and Arabic languages that the rest of AI will eventually need. A team of linguists, engineers, and native speakers begins building from the ground up.
ARCHIVED
Q2

First Language Models

~/history/2024
First ASR models validated for Hausa, Wolof, and Darija. The benchmark methodology is native-speaker evaluation panels — not just automated metrics — establishing the quality standard LocaleNLP will maintain across every language it ships.
ARCHIVED
Q3

AfriLION Architecture Locked

~/history/2024
The AfriLION transformer architecture is finalised: sub-word tokenization tuned for agglutinative morphology, tonal contour embeddings for tonal languages, and INT4-quantized inference for ARM chipsets. First offline-capable SDK tested on Tecno and Infinix hardware.
ARCHIVED
Q1

ORAOX Platform Launch

~/history/2025
ORAOX — the community data validation platform powering LocaleNLP's training pipeline — goes live. Contributors earn XP and climb leaderboards for approving, flagging, or transcribing utterances. 1,200 contributors in 60 days across 9 countries. First ORAOX-validated dataset released under CC-BY-SA.
ARCHIVED
Q3

API v1 — 20 Languages

~/history/2025
Public API launches with ASR, TTS, and translation for 20 languages. Sub-8ms latency on cloud, sub-12ms on ARM edge. 40,000 API calls in the first week from developers across Africa, the Middle East, and Europe. First enterprise pilots begin in healthcare and agri-extension.
ARCHIVED
Roadmap

50+ Languages. Edge OS.

~/history/2026
LocaleNLP targets 50+ African and Arabic language variants with state-of-the-art models for speech recognition, translation, and text processing. Edge-deployable SDK for ARM. Cultural Intelligence (CQ) API enters research phase. Expansion into B2G verticals: e-governance and identity infrastructure.
ARCHIVED
Mission

Bridging the digital language divide.

At LocaleNLP, we are building the foundational language layer for AI in low-resource environments. Not a translation API. Not a chatbot platform. The infrastructure layer — corpus collection, tokenization, acoustic modeling, edge deployment — that every African and Arabic language application will eventually need.

Our mission is to empower builders, communities, and institutions with accessible AI tools that preserve linguistic diversity and drive genuine inclusion in the global digital economy.

Community-validated data

Every dataset is collected by native speakers, reviewed by linguistic experts, and released with full provenance documentation.

Open research culture

We publish weights, benchmarks, and methodology before shipping products. Language infrastructure cannot be a black box.

Infrastructure-level thinking

Not a translation API. The corpus collection, tokenization, acoustic modeling, and edge deployment layer that everything else needs first.

Vision

A world where language is never a barrier to technology.

We envision a future where every person can access and use AI in their native language — regardless of how widely that language is spoken. Not as a niche product feature, but as baseline infrastructure.

By 2030, we aim to support over 100 African and Arabic language variants with state-of-the-art models, making digital inclusion a structural reality — not a marketing claim.

Founders

The people who started it.

AJ
Alieu Jagne
Founder & CEO
Marrakech, Morocco
Language Infrastructure · Product Strategy

Alieu founded LocaleNLP in 2024 with a single conviction: African and Arabic languages deserve the same infrastructure investment as English. He leads product vision, fundraising, and partnerships, and is the architect of the ORAOX community data ecosystem.

MG
Mouhamedou Golomanta
Co-Founder & CTO
West Africa
Systems Architecture · Edge ML · Model Training

Mouhamedou built the AfriLION training pipeline and the edge inference runtime that powers LocaleNLP models on ARM devices with zero cloud dependency. He leads all engineering, model research, and infrastructure.

50+
Languages supported
African & Arabic variants
3,200+
Community contributors
Across 6 countries
8
Active pilot countries
Healthcare, education, agriculture
4B+
Validated tokens
Community-sourced training data