No changes on our end. The VAPI custom LLM instance stopped working with streaming completions. These are our logs - And the error on VAPI logs is pipeline error. Sample Call - 65deee06-bda8-495c-a1e7-b52b4f62ecfe Error https://cdn.discordapp.com/attachments/1272646005086748812/1272646005220835478/message.txt?ex=66bbbb74&is=66ba69f4&hm=30e90d7f258907510ad7bc1d20975a3a38516732d57939d677848fca805bac19& https://cdn.discordapp.com/attachments/1272646005086748812/1272646005602386001/Screenshot_2024-08-12_at_4.00.40_PM.png?ex=66bbbb74&is=66ba69f4&hm=05c973454961a6b91b6b629db3a1fe86d41e1c351743dfdab71b40c9a6b8675f&

@Neel I'm sorry to hear that you're experiencing issues with the VAPI custom LLM instance. It seems like there might be a problem with the pipeline. To better assist you, could you please provide more details from your logs that pertain to the error? This will help me understand the specific nature of the pipeline error and guide you towards a resolution. If the error is related to one of the known pipeline errors such as

pipeline-error-playht-voice-failed

or

pipeline-error-lmnt-voice-failed

, please specify which one so I can offer more targeted assistance. If the logs contain sensitive information, please ensure to redact that before sharing.

Narrowed it down to specifically how VAPI may be now handling the custom LLM response. When we comment out cusotm LLM URL it works with openai as the provider

Hello @Neel 46, Apologies for the delayed response. We encountered an error on our end due to some changes made to our tools calling part. The issue has now been resolved, but we regret any inconvenience caused. Thank you for your understanding.