PreservingAfrican LanguagesThrough AI.
Building the world's most comprehensive open-source OCR and HTR dataset for Hausa, Igbo, Yoruba, and more — one handwritten page at a time.
What we offer
Innovative tool for
language preservation
A platform built for linguists, native speakers, and researchers to streamline the collection of high-quality African language datasets.
4 Languages
Currently supporting Hausa, Igbo, Yoruba, and Amharic languages.
Data Collection
Write and upload handwritten text samples in your native language. Every unique script helps train more accurate AI models.
Expert Verification
Review and verify submissions from contributors, ensuring quality across every dataset entry.
Fully Open Source
All datasets and models are released under open licenses. Anyone can build on our work, forever.
Real-Time Progress
Track your contributions, and watch the dataset grow.
Process
Three steps to
preserve a language
Write a page, get it reviewed by an expert, and watch your contribution improve the AI — no technical experience needed.
Write
Log in to see your active task queue. Pick up an assigned page and write the text in your natural handwriting
Upload
Scan or photograph your page and submit it. A language expert then opens your submission, reviews each line, and approves it for the dataset.
Process
Once approved, your page is added to the training pipeline thereby directly improving model accuracy.
Our Mission
Hundreds of African languages remain invisible to modern AI. We're building a handwriting dataset that will change that — openly, ethically, and in partnership with native speakers.
Join Us
Ready to make history?
Join researchers, linguists, and native speakers building the future of African language AI.