Open Research
Code for the commons.
We believe that the fundamental building blocks of African intelligence should be open. Our core models, datasets, and utilities are published here to accelerate the ecosystem.
Zamari-Labs/dara-core-sdk
Base speech synthesis pipelines for tonal African languages. Includes pre-trained phoneme maps for Yoruba, Igbo, and Hausa. Optimized for low-resource fine-tuning.
Zamari-Labs/vision-runtime
Optimized computer vision inference engine for Raspberry Pi CM4 and Jetson Nano. achieving 60fps on edge hardware for object detection tasks.
Zamari-Labs/vector-os-kernel
Hardened, education-first Linux kernel configuration. Features strict peripheral access control, kiosk-mode enforcement, and OTA update mechanisms.
Zamari-Labs/neural-bridge-protocol
Communication protocol for low-latency audio streaming between wearable devices and mobile gateways via BLE 5.3.
Zamari-Labs/yoruba-text-corpus
The largest open dataset of annotated Yoruba text for NLP training. Contains over 5M sentences with tonal markings verified by linguists.
Install via CLI
npm install @zamari-labs/core