Introduction

Aimable as your AI layer

This section is for teams building software, apps, or agentic workflows who want to run their AI through Aimable rather than calling model providers directly. Aimable becomes the governed layer between your product and the models.

The model: tenant and spaces

• Your tenant is your account on Aimable. You operate at the tenant level.
• A space is an isolated workspace inside your tenant. It has its own model policy, knowledge, tools, and governance settings.
• A space is your unit of isolation, and you decide what it represents — a customer, an environment, an application, a team. Each space isolates its own model access, data, and usage, while you keep a single integration and one set of credentials to manage.

A request is always scoped to a space, so the same backend can serve many isolated spaces without mixing their data or policies.

What you can do via the API

Everything needed to set up a space is available programmatically, under your tenant-admin credentials:

• Create a service principal and mint an API key for machine-to-machine calls.
• Create a space and set its model policy (allowed models, default, reasoning).
• Send inference through the OpenAI-compatible endpoint or the governed responses endpoint.
• Read usage and cost per space for metering and billing.

In this section

• Provisioning — create spaces, principals, and API keys via the API.
• Integration patterns — how to organize spaces, credential strategy, and model routing.
• Usage & metering — read per-space token, cost, and latency metrics.
• Evaluation — what your evaluation access includes, current limits, and how to test.

New to the API? Start with Getting Started and Integration first.