Loading IntHub…
Loading IntHub…
One OpenAI-compatible endpoint in front of 71 models from 56 providers. Prepaid in rupiah, charged per token, no subscription.
These are not marketing prices. The board reads the same table the API bills against, so the rate you see is the rate you are charged. Minimum top-up Rp 10.000.
Rate board
IDR / 1M tokens
| Model | In | Out |
|---|---|---|
| deepseek-ai/DeepSeek-V3.1 | Rp 3.000 | Rp 5.900 |
| deepseek/deepseek-v4-pro | Rp 9.100 | Rp 18.100 |
| gemini-3.5-flash-lite | Rp 3.200 | Rp 26.000 |
| MiniMax-M3 | Rp 6.300 | Rp 25.000 |
| MiniMax-M2.1 | Rp 10.400 | Rp 41.600 |
| gemini-2.5-flash-image | Rp 6.300 | Rp 52.000 |
| gemini-3.1-flash-image | Rp 10.400 | Rp 62.400 |
Most pricing pages quote their floor and stop. Here is the whole distribution instead, so you can see where the model you were planning to use actually sits before you commit a balance to it.
Rp 8.900
cheapest model
Rp 104.000
median model
Rp 624.000
most expensive
Each mark is a distinct rate in the catalogue, placed on a logarithmic axis and weighted by how many models share it. Across 71 models the dearest costs about 70× the cheapest, so the model you pick matters more to your bill than anything else on this page.
A real row from the board, priced with the same arithmetic that writes your usage ledger: ceil(in ÷ 1M × rate + out ÷ 1M × rate). You can check it against the numbers beside it.
Rp 10.000 — the smallest top-up we accept — covers 238 requests of this size.
Read live from the billing table. Search it before you sign up — no account needed to see what anything costs.
71 models
| Model | Input | Output | Copy model id |
|---|---|---|---|
| deepseek-ai/DeepSeek-V3.1DeepSeek V3.1 | Rp 3.000 | Rp 5.900 | |
| deepseek-ai/DeepSeek-V3.2DeepSeek V3.2 | Rp 3.000 | Rp 5.900 | |
| deepseek-ai/DeepSeek-V4-FlashDeepSeek V4 Flash | Rp 3.000 | Rp 5.900 | |
| deepseek-ai/DeepSeek-V4-ProDeepSeek V4 Pro | Rp 3.000 | Rp 5.900 | |
| deepseek-v3-2-volcDeepSeek-V3.2 | Rp 3.000 | Rp 5.900 | |
| deepseek-v3.1DeepSeek V3.1 Terminus | Rp 3.000 | Rp 5.900 | |
| deepseek-v3.2DeepSeek V3.2 Exp | Rp 3.000 | Rp 5.900 | |
| deepseek-v4-flashDeepSeek V4 Flash | Rp 3.000 | Rp 5.900 | |
| DeepSeek-V4-Pro | Rp 3.000 | Rp 5.900 | |
| deepseek/deepseek-v4-flashDeepSeek V4 Flash | Rp 3.000 | Rp 5.900 | |
| deepseek/deepseek-v4-proDeepSeek V4 Pro | Rp 9.100 | Rp 18.100 | |
| gemini-3.5-flash-liteGemini 3.5 Flash Lite | Rp 3.200 | Rp 26.000 |
Prices are per 1,000,000 tokens. The id in the first column is what you send as model.
The base URL and the key. Any OpenAI SDK works unchanged after that — the model id from the board goes straight into your existing call.
You get an API key immediately. It is shown once — store it.
QRIS, GoPay, DANA, or PayPal via tako.id. From Rp 10.000.
Change the base URL and the API key. Nothing else in your code changes.
curl https://api.kamiroleplay.com/api/v1/chat/completions \
-H "Authorization: Bearer $INTHUB_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"deepseek-v4-flash","messages":[{"role":"user","content":"Hello"}]}'Set your key as INTHUB_API_KEY
import os
from openai import OpenAI
client = OpenAI(
base_url="https://api.kamiroleplay.com/api/v1",
api_key=os.environ["INTHUB_API_KEY"],
)
response = client.chat.completions.create(
model="deepseek-v4-flash",
messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)Set your key as INTHUB_API_KEY
import OpenAI from "openai";
const client = new OpenAI({
baseURL: "https://api.kamiroleplay.com/api/v1",
apiKey: process.env.INTHUB_API_KEY,
});
const response = await client.chat.completions.create({
model: "deepseek-v4-flash",
messages: [{ role: "user", content: "Hello" }],
});
console.log(response.choices[0].message.content);Set your key as INTHUB_API_KEY
Top-ups go through tako.id, which accepts QRIS, GoPay, DANA, and PayPal. The minimum top-up is Rp 10.000 and the maximum is Rp 10.000.000 per invoice. There is no subscription and no card kept on file.
Per token, in rupiah. Each model has an input price and an output price per 1,000,000 tokens — the board starts at Rp 3.000 for input and Rp 5.900 for output. You are charged for the tokens you send plus the tokens the model returns, rounded up to the rupiah.
No. Your balance stays on the account until you spend it. It is only reduced by the requests you make.
Top-ups are credited to your balance and are not refundable to the original payment method. Start with the minimum top-up to try the API before you commit more.
Customer API keys are issued without a per-minute request limit. What you can send is bounded by your balance and by the upstream provider serving the model you chose, so a busy provider can return an error we pass through rather than absorb.
The board above is the live list — it is read from the same table the API bills against, so a price you see here is the price you are charged. Models are added and removed as provider availability changes, so check the list rather than hardcoding one model forever.
Create an account, top up, make your first call. No card on file, and the balance is still yours if you stop.