Hey Typebot team!I've been working on a bot that records audio within Typebot, and I use the S3 URL of that audio file to transcribe the voice via the Salad transcription API.
It was working fine until recently, but now I'm encountering an error with the Salad API.I reached out to Salad's Discord community, and their team investigated the issue.
They found that the audio isn't properly encoded and lacks the correct metadata for the Salad API algorithm to process, although the audio URL itself appears to be fine.
I believe there maybe some recent changes made to Typebot's audio feature might be causing this metadata issue. Could any Dev look into this and see if there's a solution?
