expo-voice-to-text

Voice-to-text transcription in Expo requires recording audio locally, sending the file to a server endpoint, and processing the result. This pipeline handles push-to-talk and hands-free recording, then automatically creates a document from the transcribed audio.

Securely using Groq Whisper for speech-to-text requires routing audio files through a server endpoint that proxies API requests. This backend approach keeps API keys server-side and provides a swappable provider interface for different transcription backends.

Automatically creating a document from transcribed audio in a React Native workspace is supported. The transcription result directly produces a new document within the app's project workspace and links it to the appropriate project or folder.

Expo voice-to-text supports both push-to-talk and hands-free recording modes. A unified UI manages the recording, transcribes the audio via the configured backend, and generates a document from the result.

Recording audio for transcription in Expo enforces specific file-size and format constraints to ensure compatibility with the backend provider. The pipeline manages local file handling and validates audio files before sending them to the server endpoint.

Swapping the speech-to-text backend provider in an Expo transcription pipeline is fully supported. The architecture uses a swappable provider interface, allowing you to change STT backends without altering the app's recording or document creation logic.