DuctTape.io

LLM gateway with model fallback

By DuctTape.io · published Oct 3, 2026

Create your own diagram

One gateway in front of every model: caching, routing by task, a fallback when the primary model fails, and cost tracing.