Localize Voyant Tools for
underrepresented languages

We combine human expertise with AI translation and evaluation benchmarks to build trustworthy, provenance-aware localizations.

Join the team to start reviewing AI-generated translations.

Translators

Contribute language expertise and translate UI strings into your language.

Apply

Human-in-the-loop Reviewers

Validate AI drafts, enforce placeholder and tag safety, and improve consistency.

Apply

Collaborators

Contribute datasets, glossaries, evaluation ideas, or model integrations (OpenAI / Ollama).

Learn more

Draft generation

  • NLLB-200 for multilingual machine translation drafts.
  • OpenAI (optional) for alternative draft generation and comparison.
  • Ollama (optional) for local, pluggable LLM experimentation.

Quality checks

  • XLM-R sentence embeddings for semantic similarity and back-translation checks.
  • AfroLingu-MT benchmark dataset for evaluating quality on underrepresented language pairs.