Every AI model.One balance.

One API key for hundreds of AI models, with automatic fallback across providers. Top up once and use every model.

Models
Vendors
Requests served
KENARILIVE
tokens processed
-
Today-tokens
14 daystoday
TOP MODELS TODAY
1
2
3
4
5
6
7
+17 more models

Hundreds of models. One endpoint.

See every model→

Change one line.

OpenAI and Anthropic formats both work. The rest of your code stays the same.

Read the docs→
import osfrom openai import OpenAI client = OpenAI(    base_url="https://api.openai.com/v1",    base_url="https://kenari.id/v1",    api_key=os.environ["KENARI_API_KEY"],) reply = client.chat.completions.create(    model="deepseek-v4-flash",    messages=[{"role": "user", "content": "Hello"}],)

One request, kenari finds its way.

Every request is routed through a healthy provider in priority order. If one slows down, kenari switches automatically within milliseconds. No configuration needed.

App
1 request
kenari
healthy route, automatic fallback
Provider 138ms
Provider 244ms
kenari poolstandby

Provider 1 is healthy (38ms). The request goes there.

Plans and PAYG.

Start free. Take a plan for regular use, or pay per token from your balance.

Every plan includes: connect your ChatGPT Plus/Pro or SuperGrok plan and use its allowance from any coding tool through kenari. That traffic is not billed.

PAYG

Pay per token from your balance, only when you use it. Per-model prices are in the catalog.

Solo, free

Free models included, with no subscription and no top-up.

Start Solo→
See price per model→

Start at no cost.Pay for what you use.

Sign up free and use the free models right away. Top up any time with QRIS, no credit card.