Auriko

Auriko provides an AI control plane for LLM inference, offering a model gateway, intelligent routing, observability, and FinOps in one platform. It targets teams building and running AI inference workloads who need cost-aware routing and robust reliability across multiple providers.

Auriko documents a unified API that lets applications access models and providers through a single interface, while preserving provider-specific features. It emphasizes open compatibility with existing OpenAI-style clients and frameworks, enabling minimal code changes when switching to Auriko.

Auriko highlights features like capacity-aware deployment, automatic failover, and BYOK options. It also notes that users can manage keys and routing at scale, with tools for budget controls and global edge routing to reduce latency and improve uptime.

Auriko presents itself as an infrastructure layer for AI inference, aiming to make compute liquid and priced transparently. It describes its background in quantitative trading and the goal of delivering efficient, reliable access to multiple AI models and providers.

Use cases

• Compare and select models from multiple providers in production.