Use case · Price and marketplace monitoring
Proxies for price monitoring and tracking
Track public shop and marketplace prices from the markets you serve. Choose mobile or residential traffic by the GB, and record each offer’s full context so your price tracking compares like with like.
The job
Compare the same offer, with its context.
Portproof supplies the proxy connection for your own price tracker: buy proxy traffic by the GB, then choose the pool, country and rotation in the proxy username. Payments are accepted in cryptocurrency. Your collector handles the schedule, product matching, price history and alerts. Use it for public pages you are permitted to collect; use an official feed or API when it already supplies the data you need.
A useful observation identifies the product and variant, seller, pack size, quantity, currency, displayed price, promotion conditions, stock, shipping destination, delivery charge and tax treatment. Save the source URL and capture time with those fields. An HTTP success response alone does not show that the intended offer loaded.
Hold the storefront, delivery destination, language preference, account state and browser settings consistent between checks. A country-targeted proxy supplies a network route; it does not choose the shop location, currency, tax treatment or delivery postcode for you. Record the country requested separately from anything you have verified about the result.
A small price change can hide a different total
Example only: two observations from a fictional test storefront use the same 500 ml product, seller, quantity, EUR currency and saved delivery destination. Both show the same tax treatment and available stock. Run A shows EUR 24 for the item and EUR 4 delivery: EUR 28 delivered. Run B shows EUR 25 for the item and EUR 0 delivery: EUR 25 delivered. The displayed item price rose by EUR 1 while the delivered total fell by EUR 3.
That example is comparable only because its conditions match. A different pack size needs a unit-price comparison; a different currency needs a stated conversion basis. If shipping, stock or tax treatment is missing, preserve the missing value and flag the observation for review. Do not silently turn missing data into zero.
Collection
HTTP request or browser?
Start with an ordinary HTTP request when the permitted response already contains the selected product and price. Use a browser when your task needs rendered content or a supported selection step. Verify the chosen variant and price fields before saving a row. A consent overlay, empty template or error message is not a new price.
In a browser, keep the same locale, viewport and session policy across comparable runs. Remove a resource only after checking that the fields under test still load correctly. Blocking an image may suit a text-price check; blocking the script that selects a variant can make that check wrong. Save the extraction version with the result so a parser change is distinguishable from a price change.
Use the Python Requests proxy guide for request-based collection or the Playwright proxy guide for browser setup. Then reduce browser traffic without losing needed data.
Before scaling, compare transferred traffic with usable records, including failed and repeated attempts.
Which pool
Choose the network your test needs.
Choose residential when the job calls for a peer-network connection. The pool includes home broadband and mobile peer connections; it does not guarantee a particular access network. Check the live locations list for the country you need before scheduling a run. Device availability changes, and both pools are shared.
Choose mobile 4G/5G when the test specifically needs a carrier-network exit. A mobile address does not turn a desktop browser into a phone or make an app-only offer available. Set the browser or device separately. Mobile and residential traffic draw from the same GB balance, so a controlled comparison can keep the rest of the job unchanged.
Mobile 4G/5G proxies by the GB · Residential proxies by the GB · Check countries online now
Mobile traffic runs on carrier modems operated by our network partner; residential traffic runs on opt-in peer devices in the partner network.
Settings
Keep the session aligned with the observation.
| Mode | Username | Use it for |
|---|---|---|
| Per-connection rotationFor checks that do not need a continuing proxy session. Browser cookies and local storage are separate controls; rotating the connection does not reset them. | USERNAME-peer-es-rot-ondemand | For checks that do not need a continuing proxy session. Browser cookies and local storage are separate controls; rotating the connection does not reset them. |
| Sticky sessionFor a permitted sequence that selects a store, variant or delivery destination. Keep the session name and application state together for that observation. The session keeps the same device while available; the IP may still change. | USERNAME-peer-es-sid-price1-rot-sticky | For a permitted sequence that selects a store, variant or delivery destination. Keep the session name and application state together for that observation. The session keeps the same device while available; the IP may still change. |
| Timed rotation (10-minute example)For a job designed to tolerate scheduled exit changes. A timer is not a boundary around a page or basket, so do not assume a multi-step observation will finish before the next change. | USERNAME-peer-es-rot-auto10 | For a job designed to tolerate scheduled exit changes. A timer is not a boundary around a page or basket, so do not assume a multi-step observation will finish before the next change. |
These usernames illustrate the residential pool with Spain requested; they are not a live stock promise. Replace USERNAME with your account value and choose a currently available country. Your credentials and connection builder are in the dashboard after a fulfilled purchase. Every token is listed in the docs.
A sticky session keeps the same device while available; its IP may still change. A peer device can disconnect too. If a sequence breaks, mark it incomplete and repeat it within the site’s request limits.
Compare rotating and sticky sessions · Why a sticky session can change IP
Connect your collector
Build the proxy URL in your dashboard, then use the connection format and examples. Start with the curl connection test before adding product extraction.
The echo test reports the exit address of that test request. A separate product request may use another connection; do not label its price record with the echo result as though the address was measured on the product request itself.
Cost
Plan bandwidth from a sample.
Measure a representative sample with the same browser resources, redirects and retry policy as the planned job. Record the number of complete checks and the bytes transferred, rather than assuming each product page is one network request. A browser can request additional documents, scripts, images and other resources.
Use the price monitoring traffic planner to turn a measured sample into a schedule. For planning, multiply measured bytes per complete check by checks per run and runs in the period. Count only usable observations as completed checks: keep failed attempts and their transferred bytes in the sample. If the sample already includes retries, do not add the same retries again. Use a separate allowance for uncertainty, then compare the estimate with dashboard usage after an isolated test run. Client counters and the account meter do not necessarily measure the same bytes.
Plan bandwidth from your sample and read what counts as proxy traffic.
Traffic is EUR 4.50 per GB for a small order and EUR 2.50 per GB at the largest step, on either pool. The paid trial is 0.5 GB for EUR 2.90, once per customer and subject to eligibility. View pricing and traffic validity.
Limits
Keep the collection permitted.
Collect public prices only where you are permitted to do so. Respect the site’s terms, robots.txt and request limits. Back off on a 429 or another request to slow down, and stop rather than escalating when access is refused. This offer does not cover account creation to reach restricted prices or interfering with another shop’s service.
Read the acceptable-use policy before ordering. You accept it at checkout.
Start
Test a small run before scaling.
Test the permitted workflow with a small order before scaling the schedule. The paid trial is once per customer, subject to eligibility. Its amount and price, traffic validity and order tiers are shown with the current offer. A trial is a way to test the workload; it is not a guarantee that a chosen site or country will be available.