Models

The Models page is where you add and manage every model your organisation can call through Routero. Models are grouped by the Public Model Name callers use; each group can sit over one or more deployments.

Open Models in the sidebar (organisation admins only).

The Models page

The grid view shows each model group as a card — its Public Model Name, the number of Deployments behind it, the providers, and the cost. The toolbar totals your groups and deployments. Switch to the table view to see individual deployments (the underlying model, provider, and cost per row).

The Models page — model groups as cards, with deployment counts and providers

Click a card to open the group and see every deployment behind the name.

A model group opened — the deployments, providers, and costs behind one Public Model Name

Add a model

Choose + Add Model. In the drawer:

  1. Provider — pick a provider assigned to your organisation.
  2. Model Name — choose a model, or “All {Provider} Models” to expose them all.
  3. Public Model Name — the alias your application sends to Routero (the model value in requests). Keep the default or rename it.

Click Add Model (or Test Connect first to check the provider is reachable).

Add Model drawer — provider, model name, and the public model name callers will use

No providers listed? Your platform admin hasn’t assigned a provider key to your organisation yet — ask them to add one. To bring your own provider key instead, see BYOK.

Giving several models the same Public Model Name creates a model group that Routero load-balances across — see Routing & Load Balancing.

Model settings and detail

Open a model to manage it:

  • Model Settings — per-deployment TPM/RPM limits, retries, timeout, and tags.
  • Policy — bind a policy (guardrails, prompts, memory, token saving) to the model.
  • Test Connection — verify the deployment is reachable.
  • Delete Model — remove a model you own.

Health check

From the table view, run an on-demand Health Check to see each deployment’s current status (healthy or unhealthy).

The on-demand Health Check on the Models table

→ Routing & Load Balancing — how requests spread across a model group’s deployments. → Auto Router — a virtual model that picks the best model by intent (created from Add → Auto Router). → BYOK — use your own provider key for a provider.