Keyword detection using ESP32s3 library options

[ujurmsde]

I want to inquire about possibility of detecting keywords and running model on S3 device.
I have a custom device with two microphones and I want to run a model.
I came to know about the following library but there is still not possible to create own model with custom keywords.

I know edge-impulse but I wanted to try something simpler and more effective for espressif devices that runs with IDF.

Let me know if there is some better options.

https://docs.espressif.com/projects/esp ... ation.html