Questions we actually get asked

It reaches word error rates comparable to commercial English transcription systems, which is a meaningful result for a language with as little training data as Maltese has. It significantly outperforms generic multilingual transcription on Maltese audio, particularly on code-switched Maltese-English speech.

Yes. The speech models are trained on audio from across Malta and Gozo, so they handle regional and dialectal variation. Output defaults to standard Maltese, but the system correctly transcribes Gozitan and other regional speakers.

Generic tools spread their training data across a hundred-plus languages, and Maltese gets a sliver of that. NeuroMaltese is trained and evaluated specifically on Maltese, which is why it holds up on dialect, code-switching, and domain-specific terms that trip up broad multilingual models.

Yes, in both directions. It handles documents, chatbot responses, and real-time conversation while keeping domain-specific terminology intact, which matters most in legal, government, and financial content.

Yes. Combined with NeuroDocument for OCR, it processes Maltese-language documents, including entity extraction, classification, and summarisation, which matters for Malta's bilingual government and legal work.

Yes. It's available as a managed API for speech-to-text, text-to-speech, translation, and text generation, and it integrates with other NeuroStack products for end-to-end Maltese language workflows.

Yes. NeuroMaltese is designed to slot into an existing product or workflow rather than replace it, whether that means a call centre platform, a government portal, or an internal document pipeline.