Low-resource language roadmap

Mapuzungun first. Community pathway before product claims.

OlyLive LangTech Lab is the research and deployment track for low-resource languages inside a real communication platform. The first phase centers on Mapuzungun, using corpus access, community dialogue and institutional conversations to build toward usable STT/TTS and translation tools.

Why this belongs inside OlyLive

The thesis is not philanthropic decoration. If OlyLive solves multilingual communication under real operational constraints, it can become the delivery layer for broader language access and preservation efforts.

Product realism

Low-resource language work should connect to deployed communication flows, not only offline demos or academic benchmarks.

Institutional value

Public entities, NGOs and educational systems need usable infrastructure, not merely symbolic support for inclusion.

Sustainable engine

A commercial communication platform can subsidize the cost of infrastructure, traffic and iteration required to serve underserved languages over time.

Current workstreams

The roadmap is intentionally staged, with near-term applied work and longer-term language enablement.

In progress

Mapuzungun

Preparation for a roughly 120-hour speech corpus and training-ready data pipeline aimed at real communication use cases, not only archival storage.

  • Speech collection and corpus structuring
  • Future integration path into real-time communication workflows
  • Priority focus because it is both culturally important and technically under-supported
Exploration

Aymara

Aymara is treated as the next inclusion layer for public-service and educational scenarios where language access is operationally relevant.

  • Designing applied use cases before broad model claims
  • Public-sector and NGO relevance built into the roadmap
  • Focus on usable delivery rather than only theoretical support
Roadmap

Patagonian Welsh

Patagonian Welsh represents a long-term preservation and access track where diaspora communities and institutions can benefit from product-integrated tooling.

  • Community and institutional relevance over vanity support
  • Potential voice and meeting access layer inside OlyLive
  • Built as a continuation of the low-resource language sequence

Research posture: ambitious, but careful with claims

This page should help you look more credible, not less. That means stating real progress, real conversations and real roadmap intent without pretending that exploratory work is already a finished institutional partnership.

What is already fair to say

  • There is active work toward a Mapuzungun speech corpus around 120 hours
  • There is an active community pathway in Chubut through local contacts
  • Public wording remains limited to milestones that can be disclosed without third-party confirmation

What this is not claiming yet

  • It does not imply formal signed partnerships unless they exist
  • It does not claim production-grade support before the models are ready
  • It does not confuse research readiness with commercial availability

Roadmap sequence

The path is staged so the lab can accumulate real language assets, institutional trust and product integration capacity over time.

1

Now

Develop the Mapuzungun data pipeline, define concrete use cases and keep the work tied to deployable communication scenarios.

2

Next

Extend the inclusion layer toward Aymara with institution-facing use cases in education, public service and NGO coordination.

3

Later

Move into Patagonian Welsh as a product-integrated preservation and access track for communities and institutions.

How this connects to the company model

The long-term idea is to let public institutions and NGOs finance VM capacity, traffic and inclusive deployments while the commercial product keeps the infrastructure alive and improving.

Commercial engine

Revenue from multilingual communication for organizations funds the operational base: infra, product iteration and reliability.

Inclusion engine

Institution-backed deployments create a practical path for language access and educational inclusion without separating research from real-world delivery.

Interested in collaboration, corpus strategy or institutional pilots?

Use this page as the research-facing tab of OlyLive. It is meant to signal serious intent, technical prudence and room for collaboration.

Write to hello@olylive.org