Skip to main content

Unified Orchestrator Layer

Feature RequestsChat

We need a centralized interface that seamlessly transitions between different available models without memory loss. Venice should be able to swap models on the fly to suit user needs without forgetting what it was doing to be begin with. Long term, this will be a necessity for this service as it's basically going to be standard amongst other potential services like this. 

Status: Backlog1 comment

Log in to comment and vote

Comments1

  • Amethyst Frost

    •

    Aug 4, 2025

    Hi I also wanted to add that we could do this differently than how chatgpt does it. You make the orchestrator a lightweight model that only focuses on communicating with your many open source models. Like a dispatcher. This would allows us to maintain chat memory and keep some context. This is far more doable than having a central brain that knows everything. A dispatcher ai middle man is essentially what I'm proposing. I love what Venice is and I think this is cool.