Tetelcingo Nahuatl Speech to Text: Preserving an Endangered Language with AI

Tetelcingo Nahuatl, locally called Mösiehuali̱, is a vibrant but endangered variety of Nahuatl spoken by around 1,000 people in the town of Tetelcingo, Morelos, Mexico. As a member of the Uto-Aztecan family, it shares roots with other Nahuatl dialects but has developed its own distinct phonology, including a system of vowel length and a phonemic glottal stop. Language preservation efforts are underway, but accurate documentation and transcription remain a challenge due to the lack of digital tools.

Why Accurate Speech-to-Text Matters

For a language with so few speakers, every recorded word is precious. Accurate transcription enables:

• Archiving traditional stories and oral histories for future generations.
• Creating educational materials for language classes and self-study.
• Subtitling videos on social media and community platforms, making content accessible to younger speakers and learners.
• Supporting linguistic research on dialectology and language change.
• Assisting in revitalization by providing written resources that can be printed or shared online.

Without reliable speech-to-text, many recordings remain inaccessible or require manual transcription by fluent speakers, which is time-consuming and expensive.

Challenges in Transcribing Tetelcingo Nahuatl

Transcribing this language with AI is not straightforward. Here are the main hurdles:

• Limited data – Most ASR systems are trained on large corpora for major languages. Tetelcingo Nahuatl has negligible publicly available data.
• Phonological nuances – The distinction between short and long vowels (e.g., tla vs. tlā) and the glottal stop (written as h or ʼ) are critical for meaning. A missed vowel length can change a word entirely.
• Code-switching – Speakers frequently mix Spanish words, especially for modern concepts. A good model must handle bilingual input without confusion.
• Orthographic variation – Although the community uses a Latin-based script with diacritics, older recordings may lack consistent spelling.

Speechyou tackles these issues head-on by training a custom model on Tetelcingo-specific audio, supplemented with synthetic data. The system recognizes vowel length and glottal stops with high accuracy, and it can transcribe mixed-language speech naturally.

Use Cases for Tetelcingo Nahuatl Speech-to-Text

Oral History Preservation

Elders are the keepers of traditional knowledge. Transcribing their stories ensures that this wisdom is not lost. Speechyou can process hours of recorded interviews and generate searchable text, which can be archived and annotated.

Language Revitalization

Community schools and workshops can use transcribed texts as reading materials. Subtitles for videos of cultural events make the language visible to younger generations who are more comfortable with screens.

Academic Research

Linguists and anthropologists studying the Nahuatl family or the effects of language contact can rely on Speechyou to produce accurate transcriptions of field recordings, saving weeks of manual work.

Media and Content Creation

Local radio stations, podcasters, and YouTube creators can add subtitles to their Tetelcingo Nahuatl content, reaching a broader audience and improving searchability.

How Speechyou Helps

Speechyou is designed for low-resource languages. For Tetelcingo Nahuatl, it offers:

• Real-time transcription or file upload with fast processing.
• Export to SRT and VTT for subtitles.
• Custom vocabulary to add specific terms like place names or traditional phrases.
• Bilingual support for Spanish code-switching.
• No per-minute pricing – included in the Solo plan.

By making speech-to-text accessible for endangered languages, Speechyou empowers communities to preserve their linguistic heritage in the digital age. Try it today and see how accurately it transcribes your Tetelcingo Nahuatl audio.