Feature request: audio file transcription (call center usecase)

[rhanka]

Using Danswer as support for knowledge management for a call center is a usecase we have for a client.

Custom agents are very usefull to help to have better answer based on a knowledge bank. And custom agents help to qualify knowledge base to help to define better content

I used custom agent based on RAG to reformat and enrich the transcript with annotation of documents reference, to help for triage of good customer support conversation.

So here is the request: i'd like to have support of whisper or other transcription site.

Note: this could be just seen as a tool. But every api for audio requires multipart and tools dont support multipart to send files.