Is there a way to get the LLMS (OpenAI Conversations) or any other to be able to set the fan speeds?
#assist LLM's can't set fan speed
royal stag
muted trench
Assist cannot do that - that means it doesn't expose it as tool - so neither of LLMs can do that.
You need to write custom script for that - and make it available for LLM.
royal stag
seems odd to me that the fan controls default have speed but just not via assist. Some other searching showed a possible custom intent, would this then function?
The original Extended OpenAI Conversations worked with the Functions.. I had this in the prompt:
Only use the execute_services function when smart home actions are requested.
If service fan.set_speed is to be used, use fan.set_percentage instead.
but that doesn't work in the OpenAI Conversation (Proper)
gritty turtle
have to write either a script or an intent script and expose it to assist, then a LLM can make use of it
shadow lily
seems odd to me that the fan controls default have speed but just not via assist
very, very few controls exposed in HA are available in Assist
muted trench
Yeah it's still pretty rudimentary. What is even worse, i think as there will be more things to choose from, Hassil can become slower.
shadow lily
That was dealt with about 2 months ago. It was made faster by a couple of orders of magnitude