Back to blog
Guide·May 24, 2026·5 min read

Claude API introduction for Japanese developers — everything from pricing and latency to JCB payments

A practical guide to using the Claude API from Tokyo, covering the pricing difference from Anthropic's official service, latency from Tokyo, JCB / Apple Pay / Google Pay support, Japanese token usage, and business information required under Japan's Act on Specified Commercial Transactions.

We’ve collected answers to frequently asked questions about calling the Claude API from Japan: the price difference compared with Anthropic directly, payment with JCB cards, measured latency from Tokyo, and where to find the business operator information required under the Act on Specified Commercial Transactions—everything developers should check first.

First API call in 5 seconds

Kunavo is compatible with the OpenAI wire protocol, so if you already use the OpenAI SDK, migration requires only a one-line change in base_url. In Python:

quickstart.py
from openai import OpenAI

client = OpenAI(
    api_key="sk-kn-...",
    base_url="https://api.kunavo.com/v1",  # ここを変えるだけ
)

resp = client.chat.completions.create(
    model="claude-sonnet-4-6",
    messages=[{"role": "user", "content": "日本語で挨拶してください"}],
)
print(resp.choices[0].message.content)

It’s the same for Node.js / TypeScript:

quickstart.mjs
import OpenAI from "openai";

const client = new OpenAI({
  apiKey: process.env.KUNAVO_API_KEY!,
  baseURL: "https://api.kunavo.com/v1",
});

const resp = await client.chat.completions.create({
  model: "claude-opus-4-7",
  messages: [{ role: "user", content: "東京の今日の天気を教えて" }],
});
console.log(resp.choices[0].message.content);

Registration is free. Top up from $10, pay as you go, and your balance never expires. That’s enough to try Claude Opus 4.7, Sonnet 4.6, and Haiku 4.5.

Price difference compared with Anthropic directly

Claude models on Kunavo are 30% off Anthropic’s official prices (depending on the model). Claude Sonnet 4.6 costs $3.00 input and $15.00 output per 1M tokens directly from Anthropic, compared with $2.10 input and $10.50 output on Kunavo (30% off). With prompt caching, cache reads cost 10% of the input price—that is, $0.21 per 1M tokens for rereads. Per-model pricing is available at /pricing.

Latency from Tokyo

Kunavo’s entry point is an anycast edge network, where TLS terminates at a nearby node. The gateway itself runs in a single US-East region (Ashburn, Virginia). The edge shortens the TLS handshake round trip, not the leg across the Pacific—from Tokyo, round-trip time is dominated by the physical distance between Tokyo and the US East Coast, plus the model’s own inference time. Calling Anthropic’s us-east-1 directly covers the same distance. Streaming uses an SSE pass-through implementation to provide a time-to-first-token on par with upstream.

Payment methods — JCB, Apple Pay, Google Pay

Payments are handled natively by Stripe. Available methods for customers in Japan:

  • Credit / debit cards: JCB, Visa, Mastercard, American Express
  • Apple Pay, Google Pay (when configured in Chrome / Android), Link (one-click payment)

Prices are in U.S. dollars, but Stripe automatically converts them to Japanese yen at checkout. The conversion rate includes a 2–4% conversion fee paid by the buyer (this does not apply when paying in U.S. dollars; Stripe Adaptive Pricing documentation, checked 2026-10-03). Bank transfers, convenience-store payments, and PayPay are not supported. The minimum top-up is $10 USD (approximately 1,500 yen). Because this is a prepaid wallet rather than a subscription, unused balances do not expire.

Business operator information under the Act on Specified Commercial Transactions

Business operator information, refund policy, business hours, and other details for customers in Japan are collected at /legal/tokutei-shoutorihiki. The basic refund policy is “consumed top-ups are non-refundable; unused balance is refunded when the account is closed.” Failed requests (4xx/5xx) are never charged.

Popular model choices

  • Claude Opus 4.7 — The strongest reasoning capability. For complex tasks requiring an understanding of Japanese nuance, long-context summarization, and code generation.
  • Claude Sonnet 4.6 — A balance of price and capability. The model most widely adopted by Japanese startups. Suitable as the main model for chatbots, content generation, and agents.
  • Claude Haiku 4.5 — Fast and inexpensive. For classification, filtering, simple summaries, and agent preprocessing. About one-tenth the cost of Sonnet.

Notes specific to Japanese

Claude provides the same quality for Japanese input and output as it does for English, but Japanese tends to use about twice as many tokens (approximately 0.5–0.8 tokens per character, compared with about 0.25 tokens per English character). Keep this in mind when calculating costs.

The most reliable way to specify whether to use honorific or casual Japanese is to state it explicitly in the system prompt:

“You are a customer support representative in Japan. Always respond politely using the desu/masu style, and treat customers respectfully.”

Getting started

If everything looks good so far:

  1. /app/signup Sign up for free (top up from $10, pay as you go)
  2. /app/keys Generate an API key
  3. Make your first call using a snippet from /docs/quickstart

Questions? Email contact@kunavo.com. We’ll respond in Japanese within one business day.

Related deep dives