Whisper API - Timeout While Generating Response for Recordings

Hello, folks.

Our company (not this account) is using the Whisper API to transcribe long audio files. This has yielded beneficial results and improved the quality of services, very nice.

However, we have noticed an issue. We are receiving an error from the Whisper API for one or more records. The error is as follows:

“error”: {
“message”: “Timed out generating response. Please try again with a shorter prompt or with max_tokens set to a lower value.”,
“type”: “internal_error”,
“param”: null,
“code”: “request_timeout”
}

It appears that the Whisper API (from OpenAI’s side) has timed out (The duration of the record file is approximately 1 hour.).

Do you have any suggestions?

Hi @BrianLovesAI ! I don’t think writing for OpenAI to solve this issue would be a good option.) I would suggest splitting this long files into two parts, send it as separate files, and then just combine the output in one txt file. This Python library might be a good tool for it:

I noticed the same problem for several weeks, even by limiting the chunks to 15MB instead of the 25MB limit imposed by the API, it happens and it’s random, when I retry the same request with the same file immediately after, it works perfectly GOOD