BISSARA TAUSUG AI

While tech giants spend billions training AI on English and Mandarin, Philippine indigenous languages are being completely left behind in the AI revolution. BISSARA AI solves this.

BISSARA is the first open-source, fine-tuned Large Language Model specifically engineered to understand, speak, and translate the Tausug language. Built on top of Meta's 8-Billion parameter LLaMA-3 architecture using QLoRA, BISSARA features a completely custom offline data pipeline. The training dataset was meticulously digitized, cleaned, and manually verified by a native speaker.

What makes BISSARA unique is its accessibility. The final AI weights are compressed into GGUF format, meaning any Filipino student or researcher can download the model and run it completely offline on consumer laptops using Ollama. Language belongs to the people, and BISSARA ensures the Sulu Archipelago is represented in the future of artificial intelligence.