Hey Product Hunt 👋 I started building Routor because I kept coming back to one question: **Why are we choosing one model before we even know what the request needs?** A JSON extraction, a simple support question, and a hard reasoning problem are completely different jobs, yet most AI apps either send everything to one model or make developers manually decide which model to use. I didn’t want another place that simply gives access to more models. I wanted something that could make the **model decision** for you. That became Routor. Every request is scored and classified before routing, so simple work can go to a cheaper capable model while harder work can move up to a stronger one. And we didn’t want the routing to become another black box. For every request, Routor shows you what happened: the **model, category, tier, confidence, cost, and routing decision**. You also don’t have to give Routor complete control: * **Auto**: Routor decides based on the request * **Tier**: you choose the level of capability you need * **Direct**: you choose the exact model The idea is simple: **Let the task choose the model while keeping you in control.** We’re especially building this for developers working on AI apps, agents, support bots, and workflows where requests vary a lot in difficulty. I’d really love feedback from the Product Hunt community on two things: **Would you trust automatic model routing in production?** **And what would Routor need to show you before you would?** Thanks for checking it out. I’ll be here all day answering questions and taking notes. 🙌

ROUTOR

Route by task, complexity, or exact model.

Route by task, complexity, or exact model.