I want to test whisper API (speech-to-text) before purchasing it, so I have created an API account and got $5 to use. but when I tried to use the API key then it generated the below error, so is there any testing account or something from where we can test this API before going for the billing?
message: ‘You exceeded your current quota, please check your plan and billing details. For more information on this error, read the docs: https://platform.openai.com/docs/guides/error-codes/api-errors.’,
type: ‘insufficient_quota’,
param: null,
code: ‘insufficient_quota’
But why do you want to run it via cloud API, if it’s not for a big site, you can just self host, it barely requires any CPU or RAM for the tiny models, and they’re remarkably good, in fact, the large models have not provided any benefit, except for multi language, in my tests.
There are a lot of implementations too. I’ve been using whisperlive from https://github.com/collabora/WhisperLive, but it’s for live transcription, might not have the same interface.
I got it to transcribe all the audio going out of my PC with just this:
$ pip install whisper_live
$ docker run -it --gpus all -p 9090:9090 ghcr.io/collabora/whisperlive-gpu:latest
$ python
>>> from whisper_live.client import TranscriptionClient
>>> client = TranscriptionClient("127.0.0.1",9090,lang="en",translate=False,model="small",use_vad=False)
>>> client()
This is a fancy one though… I think whisper.cpp has an OpenAI compatible API built in… if not, I’m pretty sure I can find you one in 5 minutes…
OpenAI has recently discontinued the free tier and you will need to first add credits to your account in order to use the API independent of the model used.