Prices come back in the wrong currency
Check The exit country may not match the storefront.
Fix Check the exit with ipinfo.io and use the domain of the same country.
Each marketplace runs country sites, so the price you collect depends on the country your request comes from.
Shopee, Lazada and AliExpress show prices, currency, delivery and the seller list by country or storefront. A request from the wrong country can return another storefront or a different price, and the number you save is then wrong for your market.
So set the proxy country to the market you analyse, and confirm it before the first run. For price comparison across countries keep one connection per country.
| Item | What we read | What it means for a script |
|---|---|---|
| Shopee (shopee.sg) | Other bots (*): crawl-delay 1 second; /cart/, /checkout/, /user/, /me/, /order/ disallowed | Keep to one request a second or slower and stay off account and checkout paths |
| AliExpress | All crawlers (*): /search/*, /items/*, /product/*, /productdetail/*, /api/* disallowed; some named crawlers blocked entirely | Search and product paths are off limits to crawlers by this file; look at official APIs or licensed data instead |
| Lazada (lazada.sg) | All crawlers (*): only a login-signup path and a few "undefined" paths disallowed; no crawl-delay | Few listed limits, but the terms of use and polite pacing still apply |
Sources: robots.txt of shopee.sg, aliexpress.com and lazada.sg, opened 17 November 2026. The files change; read the current one before each project. This is not legal advice.
Collect public pages only. Do not sign in with accounts that are not yours, and do not collect personal data. If you sell on the platform, its seller or open platform APIs are the supported source for your own data.
Pace requests, set a timeout, and stop if you start getting 429 or repeated checks; adding more IPs to push through is not the fix. See the 429 guide for slowing down properly.
import time
import requests
# One proxy per country, copied from My assets (HTTP port 2222, same login for HTTP and SOCKS5).
PROXIES = {
"SG": "http://LOGIN:[email protected]:2222",
"MY": "http://LOGIN:[email protected]:2222",
}
URLS = {
"SG": "https://www.lazada.sg/",
"MY": "https://www.lazada.com.my/",
}
for country, url in URLS.items():
p = {"http": PROXIES[country], "https": PROXIES[country]}
ip = requests.get("https://ipinfo.io/json", proxies=p, timeout=(5, 20)).json()
print(country, "exit country:", ip.get("country"))
r = requests.get(url, proxies=p, timeout=(5, 20))
print(url, r.status_code, len(r.text))
time.sleep(2) # keep the pace slowIllustrative example: replace LOGIN, PASSWORD and the two pages. For a SOCKS5 line, install requests[socks] and use socks5h:// with port 3333.
Open Rotating Residential Proxies. Budget is US$1.50 per GB, Premium US$5 per GB, and traffic does not expire. See the traffic guide for how many GB a price job needs.
Set Region to each market you read (Singapore, Malaysia and so on) and generate one connection each. Copy host, port, username and password from My assets.
Run the script above or open ipinfo.io. Expected: the country you set for each connection.
Fetch only the public pages you need, keep a pause between requests and save the date with every price.
Check The exit country may not match the storefront.
Fix Check the exit with ipinfo.io and use the domain of the same country.
Check Too many requests too fast, or a flagged exit.
Fix Slow down, reduce parallel requests and open a Support ticket with the order number if the exit itself looks wrong.
Check requests[socks] is not installed.
Fix Run python -m pip install "requests[socks]" or use the HTTP line on port 2222.
Rotating residential in the country of the Shopee site you read; US$1.50 per GB on the Budget line.
Its robots.txt disallows search and product paths for all crawlers, and the site terms may say more, so check both and consider official or licensed data.
No. A new IP is given on each new connection, and a reused connection keeps its IP. Pace requests rather than chasing IPs.
It depends on page size and volume; see the price monitoring traffic guide.
Official documentation · Confirmed on:
Opened on 17 November 2026: the group for other bots (*) sets a crawl-delay of 1 second and disallows paths such as /cart/, /checkout/, /user/, /me/ and /order/; named search-engine bots have a longer disallow list.
robots.txt is a crawler-etiquette file and can change; it is not a legal opinion and the site terms may say more.
Official documentation · Confirmed on:
Opened on 17 November 2026: for all crawlers (*) the file disallows /search/*, /items/*, /product/*, /productdetail/*, /api/* and account or cart paths, and blocks several named crawlers from the whole site.
robots.txt is a crawler-etiquette file and can change; it is not a legal opinion and the site terms may say more.
Official documentation · Confirmed on:
Opened on 17 November 2026: the group for all crawlers (*) disallows only a login-signup path and a few "undefined" paths, with no crawl-delay stated; separate groups exist for OAI-SearchBot, GPTBot and ChatGPT-User.
robots.txt is a crawler-etiquette file and can change; it is not a legal opinion and the site terms may say more.
Official documentation · Confirmed on:
Opened on 17 November 2026: proxies are passed as a dict to the proxies argument; user:password goes in the URL; SOCKS needs pip install requests[socks]; the socks5 scheme resolves DNS on the client and socks5h on the proxy; requests have no timeout unless one is set.
Product observation · Confirmed on:
Product facts file of the site (dated 10 November 2026), read on 17 November 2026: per-GB prices, per-IP prices and lease terms, protocols and ports, countries, no free trial.
Create an account, top up once, then pick rotating residential, static ISP or datacenter IPs and have one running in your browser within minutes.