InferCore is an open-source AI inference control plane. Route requests to vLLM, Triton, or any OpenAI-compatible backend with tenant policy, cost-aware routing, reliability orchestration, optional RAG, and Prometheus-native observability.