About the Voice Chatbot
To access the voice chatbot it is about ‘calling’ it, i.e. pressing the telephone icon.
The voice chatbot uses a cloned voice of Suzi, taken from the various files I have. There can be some modifications of speed, creativeness etc. The LLM used to present the data currently is Chatgpt 4.1 mini as it uses the least amount of gas. However, any of the main LLMs can be used e.g. Claude or Gemini.
The prompt used can be modified. It is ‘proogrammed’ to provide which Clinic or podcast episode it is quoting from. It does not give the timeslot, but this can be added if required.
About the Text Chatbot
The text chatbot is using the same data as the voice chatbot. It is interactive and provides similar answers, but typically they are shorter.
About the Interactive Avatar Chatbot
If you like, it is possible to create a custom Avatar of you (Suzi Carter); however, this requires a video to be prepared under specific conditions and costs $49 p.m. to maintain. Therefore, the example person is one of the standard avatars provided by HeyGen. Here is the link:
Currently, only limited data has been uploaded, as I need to find a suitable place to store the PDF with the clinic data, but that should be available shortly. Currently it only has 10 of the podcasts, but again, the clinic data can be accommodated.
