Developer documentation

One key, one endpoint. The Neural Gateway routes every request across Microbix Fast, Pro, and Ultra-Reasoning clusters, with failover handled for you.

Overview

The Microbix API gives you programmatic access to the same gateway that powers Aether OS and the CLI. Every request is billed transparently in mX credits, shown before you send it.

Authentication

Pass your API key as the x-aether-key header on every request. Keys are scoped per-project and can be rotated from your dashboard at any time.

// every request needs this header
headers: {
  "x-aether-key": "YOUR_KEY"
}

Chat completions

POST/v1/chat

Send a message array and get back a routed completion from Fast, Pro, or Ultra-Reasoning depending on the model field.

const res = await fetch("https://api.microbix.in/v1/chat", {
  method: "POST",
  headers: {
    "Content-Type": "application/json",
    "x-aether-key": "YOUR_KEY"
  },
  body: JSON.stringify({
    model: "microbix-pro",
    messages: [{ role: "user", content: "Hello, Aether" }]
  })
});
const data = await res.json();

Field	Type	Description
model	string	microbix-fast, microbix-pro, or microbix-ultra
messages	array	Ordered list of role/content message objects

List models

GET/v1/models

Returns the available routing tiers, current status, and per-tier mX cost estimates.

Usage

GET/v1/usage

Returns your remaining mX credits for the current billing period, plus a per-day breakdown.

Error handling

The API returns standard HTTP status codes. 401 means an invalid or missing key, 402 means you're out of mX credits, and 429 means you've hit your rate limit.

Rate limits

Free tier is limited to 20 requests/minute, Pro to 120 requests/minute, and Ultra to 600 requests/minute across a five-tier failover matrix.