En ligne

A community language-data infrastructure

A digital future for underrepresented languages.

DR Congo Starting with Lingala and the Congolese context.

Lokota helps Congolese communities document, validate, structure and govern their language resources — text, translations, varieties, voice — with an approach designed to be adapted to other underrepresented languages.

Contribute to Lokota →Explore languages

Your language isn’t on Lokota yet? Propose a language

2languages (pilot phase)
1contribution
1validated
—a growing community
A speaker records her language with headphones and a microphone, outdoors.

Our languages, our voice, our digital future.

Why Lokota

Languages spoken by millions of people stay almost invisible in the digital world.

Without reliable data, a language is missing from search, translation tools, speech recognition and AI systems. That limits access to information, education and services — and a language absent from today’s data risks being absent from tomorrow’s technology.

What guides Lokota

Community first

Speakers contribute to, validate and govern their own resources.

Consent and ethics

Clear data use, attribution, transparent reuse conditions.

Quality and diversity

Not just more data: reliable resources that represent real-world usage.

Responsible reuse

Governed dataset access for research, education and more inclusive AI.

Contribute

Three ways to grow a language

Anyone can take part, whatever their expertise. The first contribution stays simple; advanced metadata comes later.

Languages

  • Words and expressions
  • Translations
  • Regional and urban varieties
  • Contextual usage examples

Voice

  • Pronunciations
  • Audio recordings
  • Authentic speaker audio
  • Oral knowledge

Data

  • Structured resources and metadata
  • Community validation
  • Open datasets
  • Reuse for research and AI
Contribute →Review contributions

How Lokota works

From contribution to the tools of tomorrow

Lokota is not just a dictionary: it is a pipeline that turns contributions into reusable data infrastructure.

  1. Collect

    Communities submit words, phrases, translations, varieties and voice.

  2. Verify

    Contributions are independently reviewed and confirmed.

  3. Structure

    Information becomes documented, standardised data.

  4. Share

    Datasets and resources made available for research and education.

  5. Build

    Translation, voice, learning tools, multilingual assistants, research.

From language to AI

One word, documented properly, opens many possibilities

LingalaMbótethe starting point
Pronunciation/ˈᵐbóte/ · 0:03speaker audio
Meaning & contexteveryday greeting“hello”, “hi”
Translationsfr · en · sw+ varieties
Validationconfirmedby the community
Structured record→ datasetopen, documented
Machine translationSpeech recognitionText-to-speechEducational toolsLinguistic researchMultilingual AI

The application layer comes after reliable data is created — never before.

Our first implementation

Starting with the languages of the Democratic Republic of the Congo

Lingala is the pilot language. Kikongo, Tshiluba and Kiswahili follow, then priority regional languages. The method and tooling are built to be adapted, later, to other underrepresented languages.

Lake and hills landscape, Great Lakes region.
Pilot project

Lingála · ISO lin

Democratic Republic of the Congo — and diaspora

1contribution
1validated
—voice recordings

Explore Lingala →

Languages on Lokota

LingálaDRC — north-west, KinshasaPilot
KikongoDRC — Kongo CentralIn preparation
TshilubaDRC — KasaïPlanned
KiswahiliDRC — eastPlanned
200+ languages in the DRCregional languages to come

The Lokota method can later be offered to other underrepresented languages beyond the DRC.

Propose a language

A community, not a database

Every voice counts, every skill counts

Portrait of a smiling speaker outdoors.

Speakers

Share their language, their words and their knowledge.

A contributor checks entries on a laptop.

Checkers

Independently review and validate contributions.

Portrait of an elder, intergenerational transmission.

Community leaders

Guide priorities and governance for their language.

A language coordinator works in front of a map.

Linguists

Bring expertise on varieties and usage.

A researcher reviews documents and a computer.

Researchers

Reuse the data for reproducible research.

Portrait of an institutional partner.

Partners

Institutions, NGOs and universities that support and collaborate.

Governance and trust

Governance is a feature, not a legal page.

Every contribution keeps a record, a consent basis and clear reuse conditions.

  • Community validation
  • Informed consent
  • Attribution
  • Transparent contribution history
  • Recognition of language varieties
  • Ability to contest a classification
  • Independent checking
  • Responsible dataset access

Research and institutions

An open language-data infrastructure

Lokota is not only a collection platform: it is a community infrastructure that can support reproducible research and responsible technology development.

For researchers →For institutions →Propose a collaboration →
Low-resource NLPSpeech technologiesMultilingual AILanguage documentationParticipatory data governanceCommunity validationDigital inclusionResponsible AI

Resources

Go straight to Lokota’s content

Datasets

Structured, open resources with reuse conditions.

Publications

Research work and methodological notes.

Projects

Collection campaigns and ongoing collaborations.

Events

Meetups, workshops and talks.

Documentation

Contribution and validation guides.

Governance

Policies, consent and access conditions.

Statistics →Understanding AI →Resources & policies →

Join the movement

Contribute, explore, collaborate.

A community-governed infrastructure so underrepresented languages have the reliable data they need — beginning with Lingala.

Create an accountDiscover the project