Admin Console

Administration

The admin surfaces control who can use the platform, what they can reach, and how requests are governed. All of these live under the Administration group in the sidebar.

Users & API keys

Create users individually or in bulk. Each new user is issued a personal API key automatically. From the Users page you can:

Provision accounts.
Grant or revoke admin rights.
Assign a tier, or reset a user’s rate limits.

Tiers

A tier is a policy bundle: which models it can use, and its rate limits (per minute, hour, and day). Every user belongs to a tier, and new users inherit a default tier. Use tiers to separate, say, an experimentation group from production teams.

Teams

Teams are multi-member workspaces with their own tier and members. They let you track usage and enforce limits at the group level, not just per person.

Guardrails

Content guardrails screen every prompt before it routes, in three layers:

L1 · Regex — fast pattern and checksum rules to block or mask known-sensitive content (card numbers, IDs, and so on).
L2 · Policy — an LLM-based judge that checks prompts against your written policy; includes a test tool.
L3 · PII — automatic detection and masking of personal information.

Actions

For each rule you choose to allow, mask, or block. Masked content is cleaned before it ever reaches a model.

Prompt management

Define system prompts that the gateway injects into requests — platform-wide or per model — so every team gets consistent behavior and guardrails without changing their client.

Platform settings

The settings surface is the control panel for the routing layer: pricing, rate-limit defaults, providers, and the model registry (including each model's context window). Changes propagate to the gateway live, so the models and limits your users see update without a redeploy.