https://vapi.ai/ logo
Join Discord
Powered by
azure gpt-4o-mini
# support
a
v
@Alberto To assist you better, could you please provide more details about the model you're referring to? Are you looking to integrate a new voice model into the VAPI system, or are you referring to something else?
You can continue this conversation by mentioning me in the message.
Got it. Tagging @User for additional help.
a
Also i cannot connect with AZURE, i can see old support ticket saying that there are some steps to follow.. I can't find those steps :/
Ok sorted it out; We need to pass only the furst oart if the target URI. Still a problem with gpt4o-mini
Oh lol... they save but then they don't stick to the page....
Also how can i use azure from the assiistant; there is no drop down for Azure OpenAI.
@User anyone? 🙂
@User i'
it's more than a week; Can someone help me? 🙂
s
@Alberto Sorry for the late response. First, you need to add your keys in the provider-credentials section. Then, select the OpenAI model from the assistant section. After that, it will be used for all your subsequent LLM requests. Dashboard URL: https://dashboard.vapi.ai/keys Import Azure API Key: https://learn.microsoft.com/en-us/azure/api-management/azure-openai-api-from-specification
a
How does it work? Microsoft do not provide all the models
THe model works different from the openAI one, as you deploy 1 MODEL.
s
You have to use the models provided by the microsoft and available in Vapi.
a
not all of them 😦
It would break existing assitant then isn't it?
s
Sorry can you rephrase your issue.
a
e.g if i jave gpt4o and gpt4o-mini deployed on VAPI I cannot use the microsoft azure AI
as microsoft only support ONE model per deployment
Would be great if you can have a drop down only for azure, vs using OpenAI
s
The reason we don't have this is because Azure deployment model deployment names vary for different users. (updated)
a
not understanding the last sentence; What would be the difference if you've already implemented with the OpenAI? Anyway no worries, if you don't support i can understand you have better feature to implement in the roadmap
PreviousNext
Powered by