English / Nepali → Limbu, reconstructed from the documented record (no fluent speaker consulted).
Neural translator NLLB-200 · our trained model
Our fine-tuned neural model (NLLB-200, trained on the Limbu parallel corpus).
Type English or Nepali and get a Limbu draft — or translate Limbu back out. First request may take
~20–40 s while the model wakes up; after that it's quick. Strongest at into-Limbu.
A second, transparent engine: it glosses each word from the
Limbu New Testament corpus + a glossed dictionary, applies attested grammar, and shows the source of
every word (and never invents one). Useful to cross-check the neural draft above.