Scrape product prices as shoppers in each country see them

One product, four countries

This script asks for one public product page as shoppers in South Africa, the United States, the United Kingdom and Germany see it, and converts the prices to dollars to compare them. It ran as shown:

# One product page as shoppers in four countries see it: each request leaves from a checked exit in that
# country (strict: fail rather than substitute), and prices are converted to one currency to compare.
# pip install httpx    export SKRYP_API_KEY=...
import os

import httpx

API = os.environ.get("SKRYP_API_URL", "https://api.skryp.dev")
out = httpx.post(
    f"{API}/v1/compare",
    headers={"Authorization": f"Bearer {os.environ['SKRYP_API_KEY']}"},
    json={"url": "https://store.steampowered.com/app/620/Portal_2/", "countries": ["ZA", "US", "GB", "DE"], "base_currency": "USD"},
    timeout=300,
).json()

print(f"{'country':<9}{'price shown':>13}{'in USD':>9}  exit used")
for row in out["rows"]:
    shown = f"{row['price']:,.2f} {row['currency']}" if row.get("price") is not None else row["status"]
    usd = f"{row['price_usd']:.2f}" if row.get("price_usd") is not None else ""
    print(f"{row['country']:<9}{shown:>13}{usd:>9}  {row['exit']['country']} ({'verified' if row['exit']['verified'] else 'not verified'}), {row['route']}")
s = out["summary"]
print(f"\ncheapest {s['cheapest']}, dearest {s['dearest']}: {s['spread_pct']}% apart; same product in every country: {s['same_product']}")
print(f"exchange rates from {s['fx']['source']}, {s['fx']['updated']}; {out['receipt']['credits']} credits")

country    price shown   in USD  exit used
ZA           20.00 ZAR     1.22  ZA (verified), http:direct
US            1.99 USD     1.99  US (verified), http:res:us
GB            1.70 GBP     2.26  GB (verified), http:res:gb
DE            1.95 EUR     2.21  DE (verified), http:res:de

cheapest ZA, dearest GB: 85.2% apart; same product in every country: True
exchange rates from open.er-api.com, Thu, 01 Oct 2026 00:02:31 +0000; 31 credits

The same game was listed at prices that differ by more than 80% once converted, because the shop prices it per market. Each row's request left from an exit checked in that country before the page loaded; the route shows how (a plain request on Skryp's own network for South Africa, a residential IP in the country for the others). Strict mode was on: if no verified exit had been available in a country, that row would have failed instead of quietly showing another country's price.

How country proof works, and what it does not prove

1. You name the country. Skryp picks a route that leaves from an IP in it: its own network where it has one, otherwise a residential or mobile IP from that market.
2. The exit is checked before the page loads, by asking a geolocation service through that exit (ipinfo.io first, with fallbacks). An IP that is not in the country is skipped, not used.
3. Every result says where it came from. The receipt carries the exit's country, city and network and whether it was verified, so you can store the proof next to the price. See Receipts.
4. Strict mode fails rather than substitutes. Without it, Skryp may use an exit elsewhere when it cannot verify one in your country; with it, that request returns an error, at no cost.

In a test of 20 countries with strict mode on, 19 returned the page through a verified exit in the requested country[30 Sept 2026]. On 22 country-priced pages in our benchmark, the right local currency came back on 21[1 Oct 2026].

What it does not prove: a verified exit shows where the request left from, not that the site priced it correctly for that market. Sites also price by account, by cookie or by delivery address. Store the receipt with the data, check a sample by hand when a number looks wrong, and use the same_product check above to catch a country that silently redirected to a different item.

From one page to a dataset

To follow every product in a listing rather than one page, collect it once as a dataset. Skryp finds where the listing's data comes from, maps the fields you named, and keys each record by a stable value such as its URL. Then refresh it on a schedule:

# From one page to a dataset: collect every product on a listing once, refresh it on a schedule,
# and read only what changed. Re-running reuses the dataset instead of creating another.
# pip install httpx    export SKRYP_API_KEY=...
import os
from datetime import datetime, timezone

import httpx

API = os.environ.get("SKRYP_API_URL", "https://api.skryp.dev")
http = httpx.Client(base_url=f"{API}/v1", headers={"Authorization": f"Bearer {os.environ['SKRYP_API_KEY']}"}, timeout=300)
NAME = "Travel books (price example)"

ds = next((d for d in http.get("/datasets").json()["datasets"] if d["name"] == NAME), None)
how = "already existed"
if ds is None:
    made = http.post("/datasets", json={"name": NAME, "url": "https://books.toscrape.com/catalogue/category/books/travel_2/index.html",
                                        "fields": ["title", "price", "availability", "url"], "max_pages": 2}).json()
    ds, how = {"id": made["id"]}, f"created now for {made['receipt']['credits']} credits"

http.put(f"/datasets/{ds['id']}/schedule", json={"every_s": 86400})  # refresh daily; 0 turns it off
info = http.get(f"/datasets/{ds['id']}").json()
due = datetime.fromtimestamp(info["schedule"]["next_run"], timezone.utc).strftime("%d %b %Y %H:%M UTC")
print(f"dataset: {info['record_count']} records keyed by {info['key_field']} ({how})")
print(f"schedule: every {info['schedule']['every_s'] // 3600} h, next refresh {due}")

now = http.post(f"/datasets/{ds['id']}/refresh").json()
c = now["receipt"]["credits"]
print(f"refresh now: {now['added']} added, {now['changed']} changed, {now['removed']} removed "
      f"({c} credit{'' if c == 1 else 's'})")
first = http.get(f"/datasets/{ds['id']}/records", params={"limit": 2}).json()["records"]
print(first)

dataset: 11 records keyed by url (already existed)
schedule: every 24 h, next refresh 02 Oct 2026 20:52 UTC
refresh now: 0 added, 0 changed, 0 removed (1 credit)
[{'title': '1,000 Places to See Before You Die', 'price': '£26.08', 'availability': 'In stock', 'url': 'https://books.toscrape.com/catalogue/1000-places-to-see-before-you-die_1/index.html'}, {'title': 'A Summer In Europe', 'price': '£44.34', 'availability': 'In stock', 'url': 'https://books.toscrape.com/catalogue/a-summer-in-europe_458/index.html'}]

A refresh compares the new records with the stored ones by key and reports what was added, changed or removed. One that fails or stops part-way never marks records as removed: a page that did not load is not a deleted product. More on this on the Datasets page.

Refresh and change events

• Schedules: hourly, every 6 hours, daily or weekly, set per dataset from the API, the dashboard or an agent's dataset tool.
• Changes: each refresh records which fields changed on which records, with the old and new values, readable from the dataset's changes list.
• Webhooks: a monitor can post each change to your endpoint as it happens.
• Spending limits: each monitor has a credit cap and pauses itself when it is reached, and daily or monthly limits on an API key or the whole workspace stop everything else[2 Oct 2026].

What it costs

Measured on 1 October 2026, priced on the Standard plan ($49 a month for 150,000 credits):

A page through a residential IP costs 5 credits, or 10 when it needs a browser, because country pricing usually means a residential exit. Advanced unblocking costs 25 credits, only when it succeeds, and failed pages cost nothing. Every workspace gets 1,000 free credits a month, on Skryp's own network; see Pricing.

When a dashboard suits you better

If you only need competitor prices for your own shop and do not want to write code, merchant tools for price monitoring, such as Prisync or Price2Spy, give you a dashboard and alerts. Skryp is for teams that want the data itself in their own systems or agents, with proof of where each price came from.