omniroute-routing

What problem does it solve?

This Skill solves the problem of efficiently choosing among multiple AI providers and models for each request so you avoid outages, reduce cost, and improve reliability without manual switching.

Core Features & Use Cases

It lets you create and manage named provider/model “combos” that automatically distribute requests, fail over on errors, and load-balance across targets. It supports 14 routing strategies (including priority, weighted, round-robin, least-used, cost-optimized, and several context-aware and circuit-aware options) and an production-friendly auto-combo mode using 9-factor scoring. It also provides combo activation/deactivation, per-combo metrics, and routing dry-run simulation to validate decisions before sending traffic.

Use case example: When building an assistant that must stay responsive and stay within budget, define a combo with a fast low-cost model plus a higher-quality fallback, then enable cost-optimized or auto strategy so requests route to the best available target while preserving continuity during provider issues.

Quick Start

Configure OmniRoute by setting OMNIROUTE_URL and OMNIROUTE_KEY, then create your first combo named prod-auto with the desired routing strategy and provider/model targets.