Hi, would like to know if there is already the option to use the new update from OpenAI the real-time API integration for voice conversations. And which are the prices? I would also like to know if it is possible to use a custom LLM (fine tuned LLM) with the OpenAI realtime integration. Appreciate your time and congrats on the project

@zaedev Hello! I'm glad to hear about your interest in real-time API integration for voice conversations. As of my last update, I can provide you with general information, but for the most recent updates, including pricing and the availability of new features such as integration with OpenAI's real-time API, you should refer to the official VAPI documentation or contact support directly. Regarding the use of a custom LLM (Large Language Model) that has been fine-tuned, typically, APIs that support custom models do allow for integration with fine-tuned versions. However, the specifics of integrating a custom LLM with VAPI's real-time conversation capabilities would depend on the current features and limitations of the API. For the most accurate and up-to-date information on pricing, feature availability, and custom LLM integration, please visit the VAPI documentation or reach out to the support team. They will be able to provide you with detailed guidance tailored to your needs. If you have any more questions or need further assistance, feel free to ask!

How do I contact support directly? Not clear on the website as it points to go to the tag support in what seems here because it is the only link they provide 🤷‍♂️

Hi, we are currently working on this

Help please https://discord.com/channels/1211482211119796234/1293999248265707520

Are you from Vapi? Are you working on what I'm looking for? Integrating custom LLM? Another question is that it needs to make a RAG from past conversations of the same user. Getzep is a good alternative but would like to know if it is possible with Vapi. Appreciate if you can confirm this.

1. Real time API will be available from Nov1st Week 2. CustomLLM x OpenAI Realtime Integration: For complete info on this you have till Nov1st Week.

Is this in the docs how to do properly a RAG? It has to be well done not just getting some basic info. That is why I mentioned Getzep as it is pretty robust for that

we do have KB option, but you can include your custom KB as well check this doc for custom-kb https://dump.sahilsuman.me/lykfk3n6.md

Thanks. But I've fine tuned the model that is why I ask about being a custom LLM. If I ask for a RAG is for past conversations with the same user so it knows on the next conversations with the user what it was said. Will look at the code if it is possible to upload the conversation directly after the call. And this should work with the OpenAI API realtime.

For Past Conversations you can use RAG and appending system messages. If use custom llm then control is on your side, then you can hack up anyway. Yeah it will work with OpenAI Realtime API.

Thanks a lot. Will look into the docs so I can properly make you after better questions 😅