It’s available on RapidAPI: https://rapidapi.com/textclf-textclf-default/api/textclf1

Users can use TextCLF with their labeled dataset to train custom models for a variety of text classification tasks such as sentiment analysis or document classification or any other task that involves classifying with text inputs.

TextCLF is lightweight, accurate and fast to train! It will help users create custom text classification models for their specific needs.

No data is saved during training or prediction. The user can choose to either save their trained model the remote server (for faster prediction) or save the model locally (for maximum privacy) or both.

You can try it and let me know what you think and if you are happy with the accuracy!