Dear OpenAI Team and anyone interested,
I would like to submit a feature request for the ChatGPT app, specifically for the speech-to-text functionality in the normal mode (not the advanced speech mode). I frequently encounter an issue where, after recording a speech input of 2-6 minutes, the transcription fails due to a possible connection issue to the API. This results in a frustrating experience as the retry button often doesn’t resolve the issue, leading to the loss of the audio input, which I may have spent significant time and mental energy preparing.
I propose the following feature improvement:
Temporary or Semi-Permanent Audio Storage: Allow the option to store the audio recording (either temporarily or semi-permanently) upon failure, with a limit on how many recordings can be stored to avoid overloading the device’s storage.
Retry Option with Stored Recording: When the transcription fails, a retry button would appear in the message area, and when clicked, the system would reference the stored recording file and attempt the transcription again.
This feature would allow users to retry the transcription hours or even days later when the issue (likely due to network load or high user traffic) has had time to settle. It would significantly improve the user experience by preventing data loss and giving users more flexibility to retry at their convenience.
Thank you for considering this suggestion. I believe this feature would enhance the usability and reliability of the speech-to-text functionality.
Best regards,
Jeremy R.