OpenRouter alternative: pay in rupiah, QRIS or crypto

A prepaid AI API gateway compatible with OpenAI and Anthropic, topped up with QRIS or USDT/USDC. How it works and how to switch in a few minutes.

Cent AIPublished 4 min read

If you are searching for an "OpenRouter alternative" because you want to pay in rupiah, what you most likely need is an AI API gateway: one key and one base URL for many models, with a way to pay that suits developers in Indonesia. This post explains how Cent AI works as that kind of gateway, how to pay for it, and how to move existing code over.

What an AI API gateway is

An AI API gateway sits between your app and the models. Your app sends requests to one address, and the gateway forwards them to the model you name in the request body. The upside: one key, one bill, and one code format for many models.

Cent AI serves three API formats on one host:

FormatBase URLPath the client appends
OpenAI Chat Completions and Responseshttps://api.cent-ai.dev/v1/chat/completions, /responses
Anthropic Messageshttps://api.cent-ai.dev/v1/messages

So the OpenAI and Anthropic SDKs work as they are. You change only the base URL and the API key.

How to pay: QRIS and crypto

Your Cent AI balance is prepaid USD. There are two ways to add to it.

QRIS in rupiah. Pick an amount in Billing, from $1. Cent AI converts it to rupiah at the live rate with no exchange spread, locks that rate when the order is created, and shows it before you pay. Pay with any bank or e-wallet app that supports QRIS. Cent AI covers the QRIS processing fee.

USDT or USDC on Solana. Cent AI credits the amount actually received after on-chain confirmation, at 1 token to $1. If you send more, the whole amount is credited. The network fee charged by your own wallet is not covered by Cent AI, and funds sent on another network or in another token may not be recoverable.

Both methods are manual: no stored card and no automatic charges. After each top-up is credited, a receipt is emailed to you with a PDF copy.

Balance and packages

There are three kinds of balance, and every request is charged in a fixed order: promo balance, then package budget, then pay-as-you-go balance.

  • Pay-as-you-go: from QRIS or crypto top-ups (crypto network fees are yours), works with every model, and does not expire while the account exists.
  • Budget package: bought once for $1.99, $9.99 or $99, with a budget equal to its price. Each size covers a set of models, and the $99 package covers every chat model. Package budget does not expire while the account is active.
  • Promo balance: from the signup credit or referrals, expires 30 days after it is granted.

If the balance usable for a model is empty, the request is answered with HTTP 402 insufficient_balance and is not sent to the model, so nothing is charged.

How prices are calculated

Every model has four rates per million tokens: input, cache read, cache write and output. Cache read is 10% of the model's input rate, and cache write is billed at the input rate. Cost is calculated per request and rounded up once. A successful request has a minimum charge of $0.00002, and a request that fails before the model accepts it costs $0.

Every response carries usage.cost, the amount charged for that request in USD. Current rates, official prices and the actual discount per model are on the pricing page. The model list can be filtered by family on the models page.

Switch from another gateway in three steps

If your code already uses the OpenAI SDK, the move is short.

  1. Create a key on the Keys page of the dashboard. It looks like cai_<secret> and is shown only once.
  2. Change the base URL to https://api.cent-ai.dev/v1 and set the new key.
  3. Match the model names to the exact names in the GET /v1/models list. Model names differ between gateways, so this is the step people miss most often.

An example with the OpenAI SDK for Python:

from openai import OpenAI

client = OpenAI(base_url="https://api.cent-ai.dev/v1", api_key="YOUR_API_KEY")
resp = client.chat.completions.create(
    model="glm-5.3",
    messages=[{"role": "user", "content": "Hello"}],
)
print(resp.choices[0].message.content)

For the Anthropic SDK, the base URL is https://api.cent-ai.dev without /v1.

Things worth checking yourself

Every gateway has its own catalog, model naming and policies. Before you switch, compare the models you use day to day with the list on the models page, then estimate the cost of your workload with the rates on the pricing page. The rate limit on Cent AI is 600 requests per minute per key; if your app needs more, spread the traffic across several keys.

The full walkthrough from key to first request is in the Quickstart, and per-tool guides are in the docs.

Sources

  1. cent-ai.dev/docs/quickstart (opens in a new tab) · accessed 2026-11-14
  2. cent-ai.dev/docs/concepts/top-up (opens in a new tab) · accessed 2026-11-14
  3. cent-ai.dev/docs/concepts/balance (opens in a new tab) · accessed 2026-11-14
  4. cent-ai.dev/docs/concepts/pricing (opens in a new tab) · accessed 2026-11-14
ShareXThreads