AI Tooling

Cortex

The AI runtime for production workloads.

Cortex is a managed inference layer that routes requests across models and providers, with automatic fallback, semantic caching, and cost controls built in.

CloudEdge

Everything you need

Model routing

Automatic fallback across providers on failure.

Semantic caching

Cut inference costs on repeated queries.

Usage-based budgets

Hard caps per project, team, or API key.

Model routing

Automatic fallback across providers on failure.

Semantic caching

Cut inference costs on repeated queries.

Usage-based budgets

Hard caps per project, team, or API key.

Technical specifications

Supported platforms
Cloud, Edge
Category
AI Tooling
Latest version
1.9.0

Release history

  1. v1.9.0

    stable

    12/7/2026

    Added semantic cache TTL controls.

Frequently asked questions

Which model providers does Cortex support?+

All major hosted providers, plus self-hosted open-weight models via a compatible endpoint.

Which model providers does Cortex support?+

All major hosted providers, plus self-hosted open-weight models via a compatible endpoint.

Ready to try Cortex?