Is dynamic model switching based on exchange number possible in Vapi? I want to use a fast native model (gpt-4o-mini) for exchanges 1-2, then automatically switch to my custom LLM endpoint for exchange 3+.

Vapi allows you to use both native and custom LLMs, and you can dynamically assign assistants via server webhooks or Live Call Control. However, the documentation does not explicitly confirm if you can switch models mid-conversation based on the exchange number (e.g., after the 2nd exchange) without restarting the session or swapping the assistant. Could you clarify if you need this model switch to happen seamlessly within a single uninterrupted call/session, or would a brief handoff (e.g., assistant swap or session restart) be acceptable for your use case? [See docs for custom LLM integration and server events](https://docs.vapi.ai/customization/custom-llm/using-your-server) Sources: - [Connecting Your Custom LLM to Vapi: A Comprehensive Guide](https://docs.vapi.ai/customization/custom-llm/using-your-server) - [Server events](https://docs.vapi.ai/server-url/events)

I would like a seamless model provider switch based on conversation turn/exchange number without function call overhead. Current workaround: Using a function tool for handoff, but it adds 1-2s delay when the function is called.

I can help you set up a dynamic routing layer that switches models automatically after a set number of exchanges, keeping the transition seamless and low-latency. This avoids the 1–2s delay from function calls. You can track exchange count within session memory or middleware. are you managing the flow directly in Vapi or through an external orchestrator like n8n or Node? @andre_serra