Whisper openai and appgver


I want to record the voice and send it to be processed by whisper openai.

I managed to record, to send it to firebase and create an API to send it to whisper By the format is aac. This format is not readible by whispet API. Is there any easy way to do convert aac in a format like wav, mp3

No “easy” way.
You’d need to use javascript for example for the conversion. Plenty of examples if you search on google.
Or maybe use an external API that does the job on the fly…

Tks. I am using cloudconvert. I sent aac and it gives a mp3. In the meantime I will learn how to do it in java. :wink: tks.

In hindsight, I fear you may need to download some library…

Try it out by asking chatGPT, it is very good at JS.
Start your prompt with “you are a web development javascript expert” and tell it that an ‘input’ object is passed to the function. Last, you have to describe the output.

@Carlos_Marn I had a thought on this. A less technical way could be to use Make.com. I dont know for sure if this would work, but I know they have options to send and receive api calls, download files, and convert files. I made a sample flow I think could work for you (screenshot attached). The flow would start when it gets the POST request to the make.com endpoint, then it would download the file, convert it, then perform a post request to the specific endpoint. You could set it to email the file instead if that works. I dont know if this would work, but its a thought.

I have a video on using make.com to create and email pdfs if it helps: Add Workflow Automation In Your Appgyver App (SAP Build Apps) With Make.com - YouTube

1 Like