Sicilian Speech to Text: Preserving a Vibrant Language with AI

Sicilian (Sicilianu) is more than a dialect; it is a full-fledged Romance language with deep historical roots. Spoken by millions across the island of Sicily, in parts of southern Italy, and in large diaspora communities in the US, Canada, Australia, and Europe, Sicilian carries the echoes of Greek, Latin, Arabic, and Norman civilisations. Yet, like many minority languages, it faces pressure from Italian and globalisation. Accurate speech-to-text tools can help preserve and promote Sicilian in the digital age.

Why Accurate Sicilian ASR Matters

For decades, Sicilian has been excluded from most commercial speech recognition systems. Google, Amazon, and Microsoft offer no native Sicilian model. Open-source ASR like Whisper performs poorly, often confusing Sicilian with Italian or producing garbled output. This forces Sicilian speakers to default to Italian when creating captions or transcripts, erasing their linguistic identity. Speechyou changes that by delivering a dedicated, high-accuracy speech-to-text engine trained on authentic Sicilian speech.

Key Challenges in Transcribing Sicilian

• Vowel reduction: In rapid speech, unstressed vowels often disappear (e.g., 'cani' becomes 'cni'). Standard ASR models fail to reconstruct these.
• Geminate consonants: Sicilian distinguishes long and short consonants, which can change word meaning. For example:
  • 'malu' (bad) vs 'mallu' (a type of disease)
  • 'pala' (shoulder) vs 'palla' (ball)
• Dialectal vocabulary: A single concept may have multiple realisations across the island. 'Boy' can be 'carusu', 'picciottu', or 'picciriddu' depending on the area.
• Orthographic consistency: Written Sicilian uses digraphs like 'gghj' for /dʒː/ and 'cchj' for /kːj/, which are not handled by generic transcribers.

Use Cases That Benefit from Sicilian Speech-to-Text

• Cultural heritage projects: Transcribe oral narratives collected by universities and local associations. Searchable texts make it easier to analyse folklore, proverbs, and traditional knowledge.
• Independent media: Sicilian-language podcasts and YouTube channels can instantly generate SRT subtitles, improving reach and accessibility.
• Education: Teachers can create captioned resources in Sicilian for bilingual schools or language courses.
• Legal and health services: Accurate transcripts of conversations held in Sicilian ensure documentation integrity.
• Accessibility: Live captions in Sicilian at public events (theatre, conferences, religious ceremonies) help deaf participants.
• Personal use: Individuals can convert old family audio recordings into text, preserving memories.

How Speechyou Handles Sicilian

Speechyou's model is built on a diverse dataset covering major Sicilian dialects (Western, Eastern, Central, and Messinese). It uses a custom language model that includes the standard orthography conventions and a broad vocabulary. The system outputs punctuation, capitalisation, and diacritics appropriate for Sicilian. It also offers a confidence score for each segment, allowing users to review low-confidence portions. With Speechyou, you can transcribe unlimited audio (up to 2 hours per file) and export transcripts as plain text, SRT, or VTT.

Bullet Points: What Makes Speechyou Different

• Native support for Sicilian, unlike any other major ASR service
• Handles vowel elision and consonant gemination accurately
• Adapts to dialectal variation without forcing a single standard
• Generates subtitles in 100+ languages from Sicilian audio (translation optional)
• No per-minute cost; unlimited transcription included in the Solo plan
• User-friendly web interface; no technical skills required

A Future for the Sicilian Language Online

By enabling Sicilian speakers to create text and subtitles from their own speech, Speechyou contributes to the digital vitality of the language. Every transcript becomes a resource for machine translation, language learning, and archival preservation. As more content is generated in Sicilian, the ASR model improves in turn, creating a virtuous cycle. Whether you are a linguist, a podcaster, a teacher, or simply someone who wants to keep the language alive, Speechyou gives you the tools to do so efficiently and affordably.

Experience Sicilian speech-to-text today. Upload your audio and see how accurately your words are captured—in your own dialect.