Japan Proxies: Japanese Residential IPs for Scraping

A Japan residential proxy routes your request through an IP address located in Japan, so Japanese websites see domestic traffic and return the prices, stock and search results a local visitor sees. This page is about Japanese proxy IPs for data collection. If you are looking for a service that buys items from Japanese sites on your behalf, that is a proxy shopping service and a different product.

Start free with 1 GB · See pricing

Also included: automatic retries and a 99.95% success rate. 1 GB free to test, no card needed.

What a Japan proxy is and when you need one

A Japanese residential proxy makes your scraper look like a person browsing from inside Japan. Japanese sites use your IP to decide whether to serve you at all, and then which prices, stock and search results to show.

You need a Japanese IP proxy when the data only exists for Japanese visitors:

• Retail prices and stock. Rakuten Ichiba, Amazon.co.jp and Yahoo! Shopping show domestic prices, points offers and delivery dates to Japanese traffic.
• Price comparison. Kakaku.com aggregates prices from hundreds of Japanese retailers and is where Japanese shoppers check prices before buying.
• Local search results. Google.co.jp ranks results differently for a searcher in Japan than for one abroad.
• Sites that refuse foreign traffic. Some Japanese services serve nothing at all to non-Japanese IPs. The next sections cover why. Japan is one of the world's largest e-commerce markets, and its data is hard to reach from outside the country.

Prefecture and city-level data: use postal codes, not IPs

The Japan proxy targets at country level. Every request exits through a Japanese residential IP, but you can't pin it to a specific prefecture or city.

For Japanese retail, the IP is rarely what sets the region. Amazon.co.jp sets delivery dates and stock from the postal code (郵便番号) you enter. Many Rakuten and Yahoo! Shopping stores charge different shipping by prefecture, with surcharges for Hokkaido and Okinawa. Set the postal code or prefecture on the site, and a Japanese IP from anywhere in the country returns that region's result.

Use a static session while you set the postal code and load several pages, so the site sees one consistent visitor. For city-level Google.co.jp results, pass the location in the search request (Google's uule parameter).

Why Japanese sites block foreign IPs

Japanese sites geo-block harder than most. Requests from outside Japan often get a 403 error, a redirect, or a notice that the service is only available in Japan (日本国内からのみ). Three reasons drive this:

1. Fraud prevention. Many Japanese retailers and ticketing sites restrict overseas traffic to cut card fraud and resale bots.
2. Licensing. Streaming services such as TVer and ABEMA hold domestic-only rights, so they refuse foreign IPs outright.
3. Bot filtering. Foreign and datacenter IP ranges are an easy filter for automated traffic. For these sites, a Japanese residential IP is a requirement, not an optimisation. A VPN or datacenter IP registered in Japan often still fails, because the IP range is flagged as hosting. Not all Japan proxy servers get through; residential IPs on Japanese consumer networks are what these filters let in.

Handling Japanese character encoding: Shift_JIS, EUC-JP and UTF-8

Most modern Japanese sites serve UTF-8. Many older retail, government and forum sites still serve Shift_JIS or EUC-JP. A scraper that assumes UTF-8 returns 文字化け (mojibake): garbled text where Japanese characters should be.

How to handle it:

1. Read the charset from the Content-Type header, then the <meta charset> tag. Don't assume UTF-8.
2. Decode Shift_JIS as CP932. CP932 (Windows-31J) is Microsoft's extension of Shift_JIS. It covers characters like ① and ㈱ that strict Shift_JIS decoders reject.
3. Watch library defaults. Python requests falls back to ISO-8859-1 when no charset is declared, which turns Japanese text into garbage. Decode resp.content yourself.
4. Encode search terms in the site's encoding. A legacy Shift_JIS site may expect query parameters encoded as Shift_JIS, not UTF-8.
5. Normalise full-width characters. Japanese pages mix full-width (ＡＢＣ１２３) and half-width (ABC123) text. Normalise with NFKC before comparing prices or product codes. The Python example in the setup section below handles all five.

Japan residential proxy vs ISP and datacenter: which to pick

MrScraper offers residential proxies only, because that is the type that gets through Japanese geo-blocking and bot filters.

A Japan datacenter proxy is cheaper per request but fails on the sites most teams need, such as Rakuten, Amazon.co.jp and Kakaku.com. For long sessions, use a static session that holds one Japanese residential IP for up to 60 minutes. If you would rather not manage blocks, CAPTCHAs and retries yourself, Web Unblocker handles them for you.

More detail: Datacenter proxies vs residential proxies.

What people use Japan proxies for

Rakuten and Amazon.co.jp price monitoring

Rakuten Ichiba and Amazon.co.jp are Japan's two largest marketplaces. Track prices, points offers (ポイント), sellers, delivery dates and stock as a Japanese shopper sees them. Add Yahoo! Shopping for a third major marketplace. [Link to Rakuten or Mercari use-case page if one exists]

Kakaku.com price comparison data

Kakaku.com (価格.com) is Japan's dominant price-comparison site, especially for electronics and appliances. One product page lists offers from many retailers, with prices, shipping and stock. It is the fastest way to see the whole Japanese market for a product.

Google.co.jp SERP tracking

Google.co.jp results differ from Google.com. Track rankings, Shopping results and ads from a Japanese IP with hl=ja&gl=jp. Add a location parameter for city-level local results in Tokyo or Osaka.

Merchandise and collectibles pricing

Game, anime and collectibles companies track merchandise prices and resale values on Mercari, Rakuten and specialist shops. Resale prices on Mercari are a direct signal of demand and scarcity for limited releases.

Japanese market research

Western companies assessing Japan use residential IPs to see local competitors' catalogues, pricing pages, reviews and ads as Japanese customers see them. Collected data usually needs encoding and full-width normalisation before analysis.

How to set up a Japan proxy

Every request goes to one Japan proxy server endpoint, proxy.mrscraper.com, with the country set in your username. Use port 10000 for HTTP (http://) or port 10001 for SOCKS5 (socks5://). No SDK is required; any HTTP or SOCKS5 client works. The MrScraper dashboard also generates ready-made integration code for Puppeteer, Playwright, CloakBrowser and Selenium.

curl

# Check that your IP resolves to Japan
curl -x 'http://USERNAME-country-jp:PASSWORD@proxy.mrscraper.com:10000' \
  'https://ipinfo.io/json'
 
# Same check over SOCKS5 (port 10001)
curl -x 'socks5h://USERNAME-country-jp:PASSWORD@proxy.mrscraper.com:10001' \
  'https://ipinfo.io/json'
 
# Search Google.co.jp for "テレビ 4K" (UTF-8 percent-encoded)
curl -x 'http://USERNAME-country-jp:PASSWORD@proxy.mrscraper.com:10000' \
  -H 'Accept-Language: ja-JP,ja;q=0.9' \
  'https://www.google.co.jp/search?q=%E3%83%86%E3%83%AC%E3%83%93+4K&hl=ja&gl=jp'

Python

import re
import unicodedata
import requests
from urllib.parse import quote
 
proxy = "http://USERNAME-country-jp:PASSWORD@proxy.mrscraper.com:10000"
proxies = {"http": proxy, "https": proxy}
headers = {"Accept-Language": "ja-JP,ja;q=0.9"}
 
def detect_charset(resp):
    # 1. Content-Type header
    m = re.search(r"charset=([\w-]+)", resp.headers.get("Content-Type", ""), re.I)
    if m:
        return m.group(1).lower()
    # 2. <meta charset> in the first bytes of the page
    m = re.search(rb'charset=["\']?([\w-]+)', resp.content[:2048], re.I)
    return m.group(1).decode().lower() if m else "utf-8"
 
def fetch_japanese(url):
    resp = requests.get(url, proxies=proxies, headers=headers, timeout=30)
    charset = detect_charset(resp)
    # Decode Shift_JIS as CP932, which covers vendor characters like ① and ㈱
    if charset in ("shift_jis", "shift-jis", "sjis", "x-sjis"):
        charset = "cp932"
    text = resp.content.decode(charset, errors="replace")
    # Fold full-width characters (ＡＢＣ１２３) to half-width for matching
    return unicodedata.normalize("NFKC", text)
 
# UTF-8 site: Google.co.jp
html = fetch_japanese("https://www.google.co.jp/search?q=" + quote("テレビ 4K") + "&hl=ja&gl=jp")
 
# Legacy Shift_JIS site: encode the query as CP932, not UTF-8
legacy_query = quote("テレビ".encode("cp932"))
 
with open("serp_jp.html", "w", encoding="utf-8") as f:
    f.write(html)

Node.js

// npm install undici
import { ProxyAgent, fetch } from "undici";
 
const agent = new ProxyAgent(
  "http://USERNAME-country-jp:PASSWORD@proxy.mrscraper.com:10000"
);
 
const url = "https://www.google.co.jp/search?q=" +
  encodeURIComponent("テレビ 4K") + "&hl=ja&gl=jp";
 
const res = await fetch(url, {
  dispatcher: agent,
  headers: { "Accept-Language": "ja-JP,ja;q=0.9" },
});
 
// Use the declared charset; Node's TextDecoder supports shift_jis and euc-jp
const type = res.headers.get("content-type") || "";
const charset = (type.match(/charset=([^;]+)/i)?.[1] || "utf-8").trim();
const html = new TextDecoder(charset).decode(await res.arrayBuffer()).normalize("NFKC");
 
console.log(html.includes("テレビ"));

For a static session (same IP while you set a postal code or page through results), use USERNAME-country-jp-sessid-{id}-sesstime-{minutes}. Requests with the same -sessid- share one Japanese IP for up to 60 minutes (10 minutes by default). Leave both out for rotating sessions. More detail: Sticky vs rotating residential proxies.

Pricing

Japan proxies cost $1.50 per GB. You pay for the bandwidth you use, and there is no extra charge for sessions, protocols or concurrency.

Every request routes through a real consumer connection in Japan, sourced with consent. MrScraper is ISO 27001:2022 certified and SOC 2 Type II compliant; see the Trust Center.

Ready to buy Japan proxy bandwidth? See pricing or start free.

Japanese residential IPs that get through

1,496,392 Japanese residential IPs at $1.50 per GB, with HTTP or SOCKS5, unlimited concurrency and a 99.95% success rate. Test them on Rakuten, Amazon.co.jp or Kakaku.com before you commit.

Start free with 1 GB · See pricing