Choose a workload, not a model

Why AuraRoute has no model picker, and what the router does instead for coding, agents and everyday apps.

Most AI APIs start by asking you to pick a model. That choice goes stale quickly: better models ship, prices move, providers have bad days. AuraRoute asks a different question: what are you building?

Five workloads

Auto is the default. The platform decides.
Coding is tuned for code generation, debugging and software work.
Research is for longer, more careful answers: analysis and summaries.
Chat is for fast, capable answers in everyday applications.
Agent is for reliable tool calling and agentic workflows.

What the router does

Each request is scored for its workload and sent to the best available provider. Your code never names a model, so when the best choice changes, nothing in your app has to.

Your app should not need to know the model.

Try it

Create a key in the dashboard and send model: "auto". If you later want to steer a request, name a workload instead of a model.