Local LLM/GPU for AI receptionist?

[admin@adamequipment]

Hi,

Seen the details on using your own GPU/LLM rig for transcriptions but what about powering the AI receptionist?
At the moment I can only see an option for using openAI API realtime engine?

Looking at realtime call processing and KB token costs in OpenAI - it seems its going to get expensive over time to run..
Is there gong to be an option to run this on local hardware in the future - to save costs like for transcriptions?

 

[Kevin Attard Compagno]

Hi. There are no plans for this in the forseeable future. One big reason is that more complex reasoning requires the more advanced models that OpenAI provides.