whisper-web
whisper-web copied to clipboard
Feature Request: instead of "transcript.(txt|json)", name output file to the input file's name
Current functionality
The Transcript component successfully downloads transcript files named either "transcript.txt" or "transcript.json" after clicking the "Export TXT" or "Export JSON" buttons, respectively. However, that leaves the user with a directory full of files named something like transcript(n).txt where n is a positive integer (on OSX).
Feature request
For audio files loaded using the FileTile, use the input file's name to name the downloaded transcription file. For audio files loaded using the UrlTile, the input filename would be the part of the URL path after the last forward slash. For audio created with the RecordTile, use something like `recording_${new Date().toISOString() }` as the input filename.
Example: After transcribing an input file named "voice_recording_01_01_25.mp3", clicking the "Export TXT" button should download a transcript file named one of the following:
- voice_recording_01_01_25.mp3.txt
- voice_recording_01_01_25.txt
- voice_recording_01_01_25.mp3_transcript.txt
- voice_recording_01_01_25_transcript.txt Whichever file naming format is chosen should also be applied to the "Export JSON" button.