Cornish Speech to Text: Preserving and Reviving Kernewek with AI Transcription

Cornish (Kernewek) is a living Celtic language that has experienced a remarkable revival. Once spoken across Cornwall, it survived in written form and was revived in the 20th century. Today, the language is taught in schools, used in media, and spoken by a growing community. Accurate speech-to-text technology for Cornish is essential for transcribing interviews, podcasts, and educational content, helping to preserve the language for future generations.

Why Accurate Cornish Transcription Matters

For a minority language like Cornish, every piece of transcribed audio is a resource for learners and researchers. Podcasts in Cornish, such as "An Kescows" or "Kernewek Nowodhow," can be transcribed to make them accessible to learners who struggle with listening comprehension. Subtitle generation for Cornish films and videos allows the content to reach a wider audience, including those who are deaf or hard of hearing. Accurate transcription also supports the documentation of spoken Cornish, capturing dialects and variations that might otherwise be lost.

Challenges in Cornish Speech Recognition

Developing ASR for Cornish is not without difficulties:

• Phonological complexity: Cornish includes sounds like the voiceless lateral fricative (ll) and the voiceless nasal (mh, nh), which are rare in other languages.
• Mutation system: The initial consonant of a word changes based on grammatical context (e.g., "kath" (cat) becomes "gath" after certain particles). The ASR must recognize these changes.
• Limited data: There are few publicly available transcribed audio corpora for Cornish, making it hard to train models from scratch.
• Dialectal variation: The three main revived forms (Middle, Late, Unified) have different pronunciations and spellings.

Speechyou overcomes these challenges by using a custom Cornish model trained on curated datasets, including recordings from the Cornish Language Office and community sources. The model handles mutations and phonetic variations with high accuracy.

Use Cases for Cornish Speech-to-Text

• Podcast and radio transcription: Automatically transcribe weekly Cornish podcasts and radio shows for archives and learner materials.
• Film and video subtitling: Add Cornish subtitles to movies, documentaries, and YouTube videos, supporting the language's presence in digital media.
• Oral history preservation: Transcribe interviews with elderly Cornish speakers to capture the last vestiges of traditional dialects.
• Language learning: Students can record themselves speaking Cornish and get instant transcripts to check pronunciation and grammar.
• Meeting and conference notes: Cornish language events can be transcribed in real time, providing accurate records.
• Social media content: Create accessible Cornish videos with subtitles for platforms like Facebook and Instagram.

How Speechyou Helps

Speechyou is the only AI transcription service that offers dedicated support for Cornish. Our platform is easy to use: upload your audio or video file, select Cornish as the language, and receive a transcript within minutes. You can edit the transcript directly in the browser and export it as SRT, VTT, or plain text. The accuracy is over 95% for standard Kernewek Kres, and we continue to improve the model with user feedback. Whether you are a Cornish speaker, a learner, or a researcher, Speechyou provides the tools you need to transcribe and subtitle Cornish content efficiently.

Get Started Today

Ready to transcribe your Cornish audio? Sign up for Speechyou's Solo plan, which includes unlimited transcription for Cornish and 100+ other languages. No credit card required for the free trial. Join the revival of Kernewek with modern AI technology.