The Shawi Language and Its Speakers

Shawi, also known as Chayahuita, is a Cahuapanan language spoken by around 20,000 people in the Peruvian Amazon, primarily in the Loreto, San Martín, and Amazonas regions. The language is an integral part of the community's identity, used in daily conversation, traditional ceremonies, and oral storytelling. However, like many Indigenous languages, Shawi is endangered due to the increasing dominance of Spanish and the migration of younger speakers to urban areas. Preserving the language is a priority for both the Shawi people and linguists, and technology can play a key role.

Why Accurate Speech-to-Text Matters for Shawi

Accurate speech-to-text for Shawi is not just a technical curiosity; it is a tool for cultural survival. By converting spoken Shawi into written text, communities can create archives of elders' stories, produce educational materials in the native language, and generate subtitles for videos that document traditional practices. Additionally, transcription aids in language revitalization programs, allowing learners to see the written form of words they hear. Businesses and NGOs working in the region can also benefit from transcribing Shawi speech for meetings, surveys, and public health announcements.

Transcription Challenges

Transcribing Shawi presents several unique challenges. First, the language has a rich phonology with voiceless stops, nasalized vowels, and vowel harmony, which many generic ASR models struggle to distinguish. For example, the difference between /p/ and /pʰ/ can change the meaning of a word. Second, there is a severe lack of labeled training data: most public speech datasets contain only a few hours of Shawi audio, making it hard for standard models to learn the language's acoustic patterns. Third, code-switching with Spanish is common, requiring the ASR to handle two languages seamlessly. Speechyou addresses these challenges by:

• Using a custom acoustic model fine-tuned on Shawi and related Cahuapanan languages.
• Allowing users to upload their own audio-transcript pairs to improve accuracy over time.
• Offering a bilingual mode that recognizes both Shawi and Spanish in a single utterance.

Use Cases for Shawi Speech-to-Text

• Oral History Preservation: Transcribe interviews with elders to create a digital archive of Shawi myths, songs, and genealogies.
• Educational Content: Generate subtitles for videos teaching Shawi grammar, vocabulary, or traditional crafts, making them accessible to learners.
• Community Documentation: Record and transcribe community meetings, land rights discussions, and health workshops for official records.
• Accessibility: Provide real-time captions for Shawi speakers with hearing loss during public events or video calls.
• Linguistic Research: Create annotated corpora for phonetic, morphological, and syntactic analysis of Shawi.

How Speechyou Makes It Possible

Speechyou is the only speech-to-text platform that actively supports Shawi. Our AI models are trained on actual Shawi speech, including variations from different regions and age groups. The system is easy to use: upload an audio file or record directly, and within minutes you receive a transcript with timestamps. You can export the transcription as plain text or as SRT/VTT subtitles for video. For the best accuracy, we recommend using clear audio with minimal background noise. Speechyou's Solo plan includes unlimited transcription time, making it affordable for communities, researchers, and small organizations. With Speechyou, the Shawi language can be written and shared with the world, ensuring its voice is never lost.