Skip to main content
Bifrost was evaluated as an OpenAI-compatible HTTP gateway for routing AI requests from workstation tools to OpenAI, Gemini, OpenRouter, and local MLX inference. It is not part of the current stack.

​
Current path

The LiteLLM router pool is the single endpoint for local, on-estate inference. It took over Bifrost’s local-routing role. Workstation tools call local MLX directly, or they route through the router pool for the shared local fabric. These cloud providers (OpenAI, Gemini, OpenRouter) are reached directly, with each tool’s own native credentials.

​
See also