Add external (or internal) hosted STT models via API (OpenAI Whisper)

[ei23fxg]

Since there are various language models, not all of which can be supported, an external API could also be used, such as the OpenAI Whisper API, which is more or less considered a standard.

This would also make it possible to use a model for Handy via LiteLLM or similar tools, thus avoiding the need to store it multiple times on the computer.
Or directly offering support for bleeding edge models, if they offer a proxy for Whisper API.

[cjpais]

Been discussed not happening for this project. You can fork if you want this

[cjpais]

to be clear handy will offer an API endpoint for people to use the local models for other applications

it will not accept 3rd party api support for transcription.

[jammsen]

to be clear handy will offer an API endpoint for people to use the local models for other applications

it will not accept 3rd party api support for transcription.

Do you plan to provide a docker-container or setup that i can use on a system with a dgpu instead of my system where im running Handy?

[cjpais]

No

[aptdnfapt]

@dudemeister23 to be clear handy will offer an API endpoint for people to use the local models for other applications

I don't know other apps features so just being clear

Hey I really like the project that you have here I was thinking to ask this exact feature. is it going to be added anytime soon ? handy as a api server ?

[cjpais]

it will be added, when, I am not sure