Since the topic of artificial intelligence is currently in vogue, I wanted to look into it a bit more.
So far, PhonerLite already supports transcribing via Whisper. But a kind of chat was missing and still doesn't work properly.
The goal should be to be able to accept an incoming call via script and play an announcement via TTS (Text To Speech). That is already possible. In the long term, the caller's speech should be transcribed and this text passed on as input to an AI model. The AI model's response is then sent back to the caller via TTS.
But there is still a long way to go before that happens.
At the moment, in the current beta version, you can either use Google Gemini (API key is required) or an Ollama (
https://ollama.com/) locally on the network (or on the same PC). You can configure and test the corresponding settings under Options/AI.
The script has a new command called "Ask". You can also use "AskSay" and the answer will be given immediately via voice.