https://vapi.ai/ logo
Join Discord
Powered by
Azure Openai: Error in saving credentials
# support
s
Hi everyone, I'm facing issues using my Azure ChatGPT deployment with Vapi. In the attached image you can see the error I'm getting all the time. The deployment exists in Azure. What am I doing wrong? As a side note: I'm a little confused. Why doesn't Vapi require the deployment name / model name? How does it find my deployment? Best regards https://cdn.discordapp.com/attachments/1357261738311155782/1357261738617213048/grafik.png?ex=67ef8fff&is=67ee3e7f&hm=fb5e34afde813d7272d6fd4610f03bb4d3fc227e7a9871a8ee4a1cc99f347419&
v
6. Double-check your API key - Verify that your API key is correct and has the necessary permissions to access the model you're trying to use. Is there anything specific from these suggestions you'd like me to elaborate on? Sources: - [Discord Thread about Endpoint Format](https://discord.com/channels/1211482211119796234/1211483291191083018/threads/1336754090880798765) - [Discord Thread about Region Issues](https://discord.com/channels/1211482211119796234/1211483291191083018/threads/1343704609096601693) - [Discord Thread about Authentication Methods](https://discord.com/channels/1211482211119796234/1211483291191083018/threads/1339956981297057892) - [Discord Thread about Waiting Period](https://discord.com/channels/1211482211119796234/1211483291191083018/threads/1338649953102463038)
Hi there! 👋 I see you're having trouble connecting your Azure OpenAI deployment to Vapi. I'm unable to provide complete information about how Vapi identifies deployments without requiring a deployment name. However, I can definitely help troubleshoot your "Error in Saving Credentials" issue! Based on the error message in your screenshot, here are several potential solutions: 1. Wait a bit longer - If you recently created the deployment, [Azure recommends waiting at least 5 minutes](https://discord.com/channels/1211482211119796234/1211483291191083018/threads/1338649953102463038) before trying to use it with Vapi. 2. Check your endpoint URL format - The correct format should be 
openai/deployments/<deployment_name>/chat/completions?api-version
(up to the equal sign). Some users have resolved similar issues by using the endpoint without including the API version number at the end. [Try adjusting your endpoint format](https://discord.com/channels/1211482211119796234/1211483291191083018/threads/1336754090880798765). 3. Verify model availability in your region - Make sure the model you're trying to use (gpt-4-2024-11-20) is available in the swedencentral region. [Region mismatches can cause this error](https://discord.com/channels/1211482211119796234/1211483291191083018/threads/1343704609096601693). 4. Check authentication method - Microsoft recommends using the 
api-key
header instead of 
Ocp-Apim-Subscription-Key
. [This authentication issue has been reported by other users](https://discord.com/channels/1211482211119796234/1211483291191083018/threads/1339956981297057892). 5. Ensure deployment name matches model name - The deployment name in Azure should match exactly with the model name provided in the Vapi menu (like gpt-4-2024-11-20). [This is a common source of errors](https://discord.com/channels/1211482211119796234/1211483291191083018/threads/1339956981297057892).
s
Ok, it was about the endpoint URL. Now it's fixed. Thanks
Well now the model is configured but the assistant crashes instantly. I suppose it can't reach the model...
s
checking if this is resolved/solved for you?
m
@sctof can you tell me where did you set the azure api keys? i can't find any place in the vapi dashboard where to set my api keys
2 Views
PreviousNext
Powered by