@Sahil can you help?

also things like "name", "parameters" and everything shouldn't be in quotes. the actual values you assign to eeach param should be in quotes but not the param name

you can't copy paste an assistants config from the vapi GET request. it shows you the general structure of everything but for the websdk at least, not everything is identical

same deal for transcriber. i don't know what environment you're doing this in but you don't give the params quotation marks. im using the widget in the websdk to setup by agent, keep in mind your specific usecase may be a bit different https://cdn.discordapp.com/attachments/1263844203688104077/1263893639059341415/Screenshot_2024-07-19_at_12.20.33_PM.png?ex=669be42d&is=669a92ad&hm=37f0e7e8933b0de177c5f344a254ae42127eeca7df252ba6d435fad6ea32de29&

can you paste your function here so that I can compare? It seems like I messed up the structure

By function, do you mean the function tool for the LLM? Or the actual functions I used to utilize it

and i sent it. my original JS config and one that should work for your python. only real differences are the booleans (python uses a capital first letter while JS does not) and the comments