One key · Every model

One project key.
Route every model reliably.

The rain that awakens intelligence.

A scoped API key per project for OpenAI-compatible, Anthropic, and Gemini-style requests, with usage logs and credit controls in one console.

How ATP Works

Create a Project

Set up an organization, workspace, and project for the app you are building.

Enable Models

Choose which model IDs the project is allowed to call.

Use the Gateway

Call /v1/chat/completions, /v1/messages, or Gemini-compatible routes with the project key.

Energy in motion

Know which key reaches which models, what it spent, and which request failed.

⚡

SCOPE Project control

Each API key belongs to a project and only uses the models enabled there.

↔

ROUTE One gateway

Use one gateway for OpenAI-compatible, Anthropic, and Gemini-style requests.

◆

OBSERVE Full visibility

Review request logs, token usage, key labels, and project credit balance.

What you get

Create keys under a project so access, logs, and credit usage stay separated.

Organize work by organization, workspace, and project before issuing keys.

Enable or restrict model IDs at the project level before traffic reaches providers.

Inspect request status, model, API key label, input tokens, and output tokens.

Create an account, make a project, then issue a scoped API key.