I am trying to make a custom gpt which can run from a user-supplied mp4 or wav files of a meeting and answer requests related to that file. Whether in a GPT or using chatGPT 4o in a browser, I get a message (after it attempts to run) that the necessary capabilities are not available in this environment. It sometime tries importing openai and sometimes whisper but can’t find them. Someone else’s GPT called Whisper Transcriber is having the same problem. I know how to transcribe locally but want this to be internal to the customGPT and don’t want to make an api call for transcription from the GPT,
Hi @tevslin , from what I know currently that’s the only option - to make a separate API call for transcription.