Hello guys, I switch from RetellAI to Vapi. On both, i have the same prompt with the same KB with the same LLM (gpt 4o). But the results are better on RetellAI (for our expectations) Do you know why ? Do we need to remake a prompt starting from Vapi ?

@RealKhalass Hello and welcome to VAPI! It's great to hear that you're exploring our Voice API for your voice development needs. Transitioning between different platforms like RetellAI and VAPI can indeed sometimes yield different results, even with the same underlying language model like GPT-4. The variation in performance could be due to several factors, such as differences in the integration setup, the way each platform handles context management, or even subtle differences in the way prompts are processed. It's not necessarily required to remake your prompt from scratch, but it might be beneficial to fine-tune it to better align with VAPI's specific features and capabilities. If you'd like, I can assist you in optimizing your prompt for VAPI. Could you provide me with more details about your current prompt and the specific expectations you have? This way, we can ensure that you get the most out of VAPI's functionalities.

@User Any ideas guys please ?

@RealKhalass Hello and welcome to VAPI! I'm glad you're exploring our services. Transitioning from one platform to another can sometimes require adjustments to optimize performance. It's possible that the differences you're noticing are due to the way VAPI integrates its audio-text model to manage turn-taking and response timing. While the underlying language model (GPT-4) is the same, the way each platform handles input and output could be different. This can affect how the prompts are interpreted and the responses generated. It might be beneficial to tailor your prompts specifically for VAPI to take full advantage of its unique features, such as its fusion audio-text model. If you'd like, I can assist you in optimizing your prompts for VAPI. Would you like to proceed with that?

Hey @RealKhalass To help track down this issue, could you share: - The call ID - When exactly this happened (the timestamp) - What response you expected to get - What response you actually got instead This would really help me figure out what went wrong!

Hey, thank for your help, actually it was my fault (I didn't check the temperature of the LLM). But I have another request "Overpriced call" could you help me please on this one ? Thank you !

What do you mean by overpriced?

Quote : Hey guy, this call looks very overpriced : 2cf81a39-ec3c-4260-a9a7-db9a5e45dea6 $3,7 for 13min with gpt 4o and 11labs (28cts/min). This call with the same setting is 20cts/min : c6a70933-5374-4c00-94cd-d4fe7b9d3938 Both are higher than the annonced price (15cts/min)

@RealKhalass I believe this was cleared in another thread, if not then do let me know.