KAEONIX
SIGN INGET STARTED

API reference

Avatar over an OpenAI-compatible HTTP API — the endpoint, the auth header, and what a request costs.

UPDATED29 JUL 2026v1.0.0

Endpoint

The API is OpenAI-compatible, so any client that speaks that protocol works by changing two settings: the base URL and the key. There is no Kaeonix-specific SDK to install.

BASE URL
https://api.kaeonix.com/v1

Models are named provider:model. Avatar is feyncode:avatar-1.0 — note the provider segment, which names the issuing company rather than the product and is a stable wire identifier.

REQUEST
curl https://api.kaeonix.com/v1/chat/completions \
-H "authorization: Bearer $KAEONIX_API_KEY" \
-H "content-type: application/json" \
-d '{
"model": "feyncode:avatar-1.0",
"messages": [{"role": "user", "content": "explain this stack trace"}]
}'

Authentication

Every request carries a bearer key issued from your account. A key is shown once at creation and stored hashed, so it cannot be recovered — rotate it rather than hunting for it.

What it costs

Requests are metered per token, in three classes, and billed against your balance in the same units the CLI reports. Current rates are on the pricing page, which reads them from the same configuration the meter bills from — the published price and the charged price are the same value.

Cached input is roughly a tenth the price of fresh input, so the shape of a workload matters as much as its size: a long stable prefix reused across many turns costs far less than the same tokens sent fresh each time.