feat: upload and stop transcription updates (!9) · Merge requests · VISWAM / apps / Speech / Voice App Frontend

Vaishnavi requested to merge feat/upload into develop Apr 23, 2026

Summary

This MR adds the audio upload workflow along with a Stop Transcription option during active processing.

Users can now upload supported audio files for transcription, and while transcription is running, a Stop button is shown to cancel the process immediately.

Changes Made

Upload Audio Feature

Added upload file area with drag-and-drop support
Added click-to-browse file picker
Added supported file type validation
Added file processing flow using existing ASR pipeline

Stop Transcription Feature

Added Stop button during active transcription
Supports stopping transcription for uploaded files
Supports stopping transcription for corpus record transcription flow
Prevents further transcript text after stop

closes #9 (closed)

Resets UI back to ready state

UI / UX Improvements

Improved upload box messaging and states
Added processing/loading feedback
Better transcription control experience

Code / Logic Updates

Added transcription state handling (isTranscribing)
Added stop/cancel logic for active worker/model processing
Added cleanup/reset flow after stopping

Supported File Types

MP3
WAV
M4A
WebM
MP4

Testing Done

Manual Testing

Uploaded valid audio files and verified transcription starts
Verified Stop button appears during transcription
Verified clicking Stop cancels transcription
Verified no further text is appended after stop
Verified UI resets correctly
Verified corpus transcription stop flow works

Regression Checks

Push-to-talk still works
Continuous VAD still works
Existing transcript actions remain functional

Screenshots / Demo

Checklist

Feature implemented
Tested locally
No known regressions
Ready for review

Notes

This MR improves usability by giving users more control during long-running transcription flows and adds a complete upload-based transcription experience.