Why we built ForroVivo: Preserving language in the age of foundation models

The digital fragility of Creole languages

Forro (also known as Santomean Creole) is an expressive Portuguese-based Creole language native to the islands of São Tomé and Príncipe in the Gulf of Guinea. For centuries, it has served as the heartbeat of oral storytelling, folklore, communal trade, and family identity.

As generations migrate and daily communication shifts into digital messaging apps, the lack of standardised digital lexicons and spell-check engines accelerates language attrition. If a language cannot be typed effortlessly on a smartphone keyboard or recognized by a search query, it faces an invisible obsolescence in modern youth culture.

“A culture does not disappear when people stop speaking its words; it begins disappearing when modern machines cannot understand them.”

Designing ForroVivo for everyday pocket use

Rather than building an academic repository hidden behind university paywalls, we designed ForroVivo as an intuitive iOS application. From the first sketch, our benchmark was accessibility: a child in São Tomé, an elder in Lisbon, or a diaspora researcher in Paris should be able to look up an idiom, hear its phonetics, and understand its cultural etymology in seconds.

We prioritized offline-first synchronization, high-fidelity audio recordings from native speakers, and an editorial design that treats the language with dignity, elegance, and cultural reverence.

Toward ForroLLM

The data structures curated through ForroVivo now power our experimental lab initiative, ForroLLM. By systematically tokenizing Creole grammar and building bidirectional translation pathways, we are demonstrating that small, purpose-built edge models can preserve cultural heritage far more faithfully than generalist trillion-parameter systems.