This creates a basic wrapper for the OpenAI API. It is not meant to be used in production, but instead locally hosted. This does not intend to provide a full replacement for OpenAI, but simply to allow same interface for local usage of Amica.
It currently provides the following routes:
v1/audio/transcriptionsv1/audio/speech
Configuration is done by editing server.py file directly.
Instead of models or voices being dynamic, all responses use what is defined in server.py.
NOTE: ensure you are using python 3.11
open terminal and run:
git clone https://github.com/semperai/basic-openai-api-wrapper.git
cd basic-openai-api-wrapper
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txtopen terminal and run:
cd basic-openai-api-wrapper
source venv/bin/activate
flask runthis will start server on port 5000
Set speaker.wav with xtts_2 model to clone voice.
Set whisper_model to larger model for better results.