Home › The Price of a Gigabyte › Proxy Pricing Transparency Report
✓ data: 12 providers’ pricing, terms, refund and KYC pages saved 2026-10-02; price history from the Wayback Machine, July 2024 to September 2026 · rubric · method · download the data
How honest is residential proxy pricing? We saved the residential pricing pages, terms, refund policies and KYC pages of 12 providers on 2 October 2026. Then we checked eight things a buyer should be able to see before paying: is there a per-GB price for every plan, is tax mentioned, what happens to unused traffic, are refund and ID-verification terms published, can you get the advertised price without a coupon code, are the struck-through “regular” prices real, and does each plan say whether it renews.
For the struck-through prices we went back to the Wayback Machine. We took the first capture of each month for each provider’s pricing page, July 2024 to September 2026, and checked whether the “regular” price had ever been the price you actually paid. The scorecard is ProxyPicker’s assessment, against a published rubric of yes/no criteria. Every mark is backed by a saved page and, where it’s a pass, an exact quote. It replaces our Pricing Transparency Audit (August 2026), which graded how easy it is to find each provider’s real rate; the old URL now redirects here. This report adds the archive check on struck-through prices, plus tax, unused-traffic, refund and KYC terms. The two use different criteria and dates, so their grades aren’t directly comparable (what changed).
- providers 12
- show a struck-through price 5
- price needs a coupon code 2
- state tax on the pricing page 4
- full marks 1
The headline numbers
- 5 of the 12 providers show a struck-through “regular” price on their residential pricing page (Bright Data, Webshare, NodeMaven, Proxy-Cheap, ProxyEmpire). For 2 of those 5 we could confirm the struck price was real: it was charged before the sale (ProxyEmpire), or it is the price you pay without the coupon (Bright Data). For Webshare it was real only on the 1 GB plan. For NodeMaven and Proxy-Cheap the archives can’t tell us.
- Webshare strikes through a $7.00/GB “regular” price on all 9 of its plans, but $7.00 was only ever the price of the 1 GB plan. In 10 monthly captures with an undiscounted price table (July 2024 to May 2025), the 3,000 GB plan was never shown undiscounted at more than $4.50/GB. The page says “SAVE 80%” on that plan; against $4.50 the $1.40 price is 69% off. On the 100 GB plan it says “SAVE 67%”; against the highest price we found ($5.50) it’s 59%. Until March 2026 Webshare struck through each plan’s own earlier price; the $7.00-everywhere version first appears in the 8 April 2026 capture.
- Bright Data has shown a coupon code on its residential pricing page in all 20 monthly captures since February 2025. Today every plan price is quoted after the 50% code: “$8 $4.00 / GB … Use this coupon code: RESIGB50”. The same RESIGB50 code was already there in August 2025. Webshare’s discount has been on its page in all 16 monthly captures since April 2025, and ProxyEmpire’s in all 9 since November 2025. ProxyEmpire’s page currently says “Limited-time — 50% OFF rotating residential proxy plans. Use code EMPIRE50 at checkout.”
- Only 4 of 12 pricing pages say whether tax is included (Oxylabs, Decodo, SOAX, Proxy-Cheap). Four others mention tax only in their terms: Bright Data (“Prices are net of any withholding or other taxes”), IPRoyal (“Services are indicated net of any withholding taxes”), NodeMaven (“Sales tax will be added to the price of purchases as deemed required by us”) and ProxyEmpire, whose terms say listed prices “do not include VAT or other taxes, unless indicated otherwise”.
- 5 of 12 pricing pages don’t say what happens to traffic you don’t use (Oxylabs, Decodo, Webshare, Rayobyte, Proxy-Cheap). Oxylabs’ self-service terms say traffic “shall be valid for 1 (one) calendar month after their purchase”; that isn’t on the pricing page. At the other end, IPRoyal and DataImpulse say their traffic never expires, and SOAX gives 60-day credit rollover.
- All 12 publish refund terms, and they range from none to 14 days. Bright Data says “The Subscription Fees are non-cancelable and non-refundable”, and ProxyEmpire says “All sales and charges are final”. At the other end, Decodo offers 14 days (first purchase, conditions apply) and DataImpulse advertises a 7-day refund policy. Proxy-Cheap refunds a balance “only if the account is permanently closed”.
- All 12 publish their ID-verification (KYC) rules, and they differ a lot. NodeMaven says “At the moment, KYC is not required”. ProxyEmpire says you don’t need it “for the trial or for normal plan purchases”. Proxy-Cheap asks for ID once you top up $300 or more or your top-ups total $250. Webshare’s compliance policy says “we require users to verify their identities”.
- On our rubric, SOAX meets all 8 criteria; Webshare and ProxyEmpire meet 5 of 8. Every provider shows a per-GB price for every plan and publishes refund and KYC terms. The most common gaps are tax (8 of 12) and unused-traffic terms (5 of 12).
The rubric
Each criterion is a yes/no fact that anyone can check on the saved pages. The score is the number of criteria met divided by the number we could assess. We don’t weight criteria and we don’t judge whether a price is good value. C1, C2, C3, C6 and C8 look only at the residential pricing page, including its on-page FAQ and anything shown by clicking a plan tab, slider or “see all plans” button, but not at checkout after entering an email or payment details. C4 and C5 count any public page on the provider’s site or help centre.
| # | Criterion | Passes when… | Providers passing |
|---|---|---|---|
| C1 | Per-GB price for every plan | Every residential plan listed shows a price per GB, visible without logging in or entering payment details. | 12 of 12 |
| C2 | Tax treatment on the pricing page | The pricing page (including its on-page FAQ) says whether prices include or exclude VAT/sales tax. | 4 of 12 |
| C3 | Unused-traffic terms on the pricing page | The pricing page says when purchased traffic expires, how long it is valid, or whether it rolls over. | 7 of 12 |
| C4 | Refund terms published | A public page states the refund terms. An explicit “no refunds” counts. | 12 of 12 |
| C5 | ID-verification (KYC) terms published | A public page states whether and when identity verification is required. An explicit “not required” counts. | 12 of 12 |
| C6 | Headline price without a code | The prices shown can be paid without entering a coupon code. | 10 of 12 |
| C7 | Struck-through prices check out | Every struck-through “regular” price is either the price the page says applies without a code, or a price shown undiscounted for that tier in a monthly Wayback Machine capture from the previous 24 months. Passes if no struck price is shown; N/A if no readable capture shows the tier undiscounted. | 9 of 12 (2 N/A) |
| C8 | Billing basis stated for every plan | The pricing page says, for every plan, whether it is a one-off purchase or a recurring subscription. | 11 of 12 |
The scorecard
✓ = meets the criterion, ✗ = doesn’t, N/A = couldn’t be checked. The note and evidence behind each mark are in the scorecard CSV, and the quotes are in the evidence CSV.
| Provider | C1 | C2 | C3 | C4 | C5 | C6 | C7 | C8 | Score |
|---|---|---|---|---|---|---|---|---|---|
| SOAX | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | 8/8 |
| DataImpulse | ✓ | ✗ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | 7/8 |
| Decodo | ✓ | ✓ | ✗ | ✓ | ✓ | ✓ | ✓ | ✓ | 7/8 |
| IPRoyal | ✓ | ✗ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | 7/8 |
| Oxylabs | ✓ | ✓ | ✗ | ✓ | ✓ | ✓ | ✓ | ✓ | 7/8 |
| Proxy-Seller | ✓ | ✗ | ✓ | ✓ | ✓ | ✓ | ✓ | ✓ | 7/8 |
| NodeMaven | ✓ | ✗ | ✓ | ✓ | ✓ | ✓ | N/A | ✓ | 6/7 |
| Proxy-Cheap | ✓ | ✓ | ✗ | ✓ | ✓ | ✓ | N/A | ✓ | 6/7 |
| Bright Data | ✓ | ✗ | ✓ | ✓ | ✓ | ✗ | ✓ | ✓ | 6/8 |
| Rayobyte | ✓ | ✗ | ✗ | ✓ | ✓ | ✓ | ✓ | ✓ | 6/8 |
| ProxyEmpire | ✓ | ✗ | ✓ | ✓ | ✓ | ✗ | ✓ | ✗ | 5/8 |
| Webshare | ✓ | ✗ | ✗ | ✓ | ✓ | ✓ | ✗ | ✓ | 5/8 |
What we found, provider by provider
Each item links to the page we saved. Where a provider fails a criterion because something is missing, we searched the saved pricing page for the relevant words (VAT, tax, expire, roll over, valid for, coupon, billed, monthly, subscription…) and found nothing relevant; those searches are in the absence-checks CSV.
Struck-through prices: how long, and are they real?
A struck-through price tells you the discount is worth something only if the “regular” price is one people actually paid. Here is what the archives show for the 5 providers that strike a price through today:
- ProxyEmpire: real. Its struck $7/GB pay-as-you-go price is the price it charged undiscounted in June and October 2025; the other plans match too (Startup 60 GB: $5.00/GB then, struck $5/GB now). The 50% discount began as a PROMO50 code in November 2025 and turned into a struck price in April 2026.
- Bright Data: real in the sense that the struck price is what you pay without the coupon: the page says to “use this coupon code: RESIGB50” for the lower price. Before the codes, pay-as-you-go was $8.40/GB in December 2024. Codes have been on the page since February 2025 (RESI50, then RESI40, then RESIGB50).
- Webshare: real only for the 1 GB plan. The table below compares the $7.00 struck on every plan with the highest undiscounted price we found for that plan. Webshare’s discount began in spring 2025 with each plan’s own price struck through (July 2025, March 2026); by April 2026 every plan showed $7.00 struck through. The prices you pay haven’t changed since July 2025. (Our August 2026 audit, which used different checks, passed Webshare on discount framing.)
- NodeMaven: can’t verify. Its calculator shows a “-20%” struck price at every plan size. Since May 2026 the calculator doesn’t render in Wayback captures. Earlier captures show different plans with coupon codes (NM80, NM30), and 2025 captures show prices in euros.
- Proxy-Cheap: can’t verify. Its rotating residential tab strikes through $2.99 and $2.60. Archived copies of its pricing page from 2024 to 2026 show only “starts from” figures, not per-plan prices.
| Webshare plan | Struck price on the page | You pay | Saving the page claims | Highest undiscounted price we found (Jul 2024–Sep 2026) | Saving vs that price |
|---|---|---|---|---|---|
| 1 GB | $7.00/GB | $3.50/GB | 50% | $7.00/GB | 50% |
| 10 GB | $7.00/GB | $2.75/GB | 60% | $6.00/GB | 54% |
| 25 GB | $7.00/GB | $2.60/GB | 62% | $5.50/GB | 53% |
| 50 GB | $7.00/GB | $2.45/GB | 65% | $5.50/GB | 55% |
| 100 GB | $7.00/GB | $2.25/GB | 67% | $5.50/GB | 59% |
| 250 GB | $7.00/GB | $2.00/GB | 71% | $5.25/GB | 62% |
| 500 GB | $7.00/GB | $1.75/GB | 75% | $5.00/GB | 65% |
| 1,000 GB | $7.00/GB | $1.50/GB | 78% | $4.50/GB | 67% |
| 3,000 GB | $7.00/GB | $1.40/GB | 80% | $4.50/GB | 69% |
The undiscounted prices come from 10 captures, from 2024-07-15 to 2025-05-08; no later capture shows an undiscounted table. Monthly sampling can miss a price that was live for only a few days. Two other providers ran codes that have since ended: Oxylabs (OXYLABS50, February 2025 to February 2026) and Decodo (RESI50, until February 2026). Rayobyte showed struck-through sale prices from December 2024 to March 2025.
What happened to the August 2026 audit
Note, 2 October 2026: our earlier Pricing Transparency Audit (pages read 7 August 2026, formerly at /pricing-transparency/) now redirects to this report. It graded 15 of 18 providers on five pass/fail criteria: full tier table, machine-readable pricing, internal consistency, honest discount framing, and headline conditions stated. We retired it instead of regrading it, because its grades and this report’s can’t be reconciled on one scorecard:
- Webshare got an A (5/5) in August, including a pass on discount framing. This report’s archive check found its $7.00 struck-through price is real only for the 1 GB plan, so it fails C7 and scores 5/8.
- Bright Data failed discount framing in August because “50% off” was shown on every check. Here its struck price passes C7 (it is the price without the code), but it fails C6 because every price shown assumes coupon RESIGB50.
- NodeMaven and Proxy-Cheap failed discount framing in August (both graded D). Here C7 is N/A for both: the archives can’t show whether their struck prices were ever charged.
- Oxylabs, IPRoyal, Proxy-Seller, ProxyEmpire, NodeMaven and Proxy-Cheap failed “full tier table” in August. On 2 October all 12 providers here showed a per-GB price for every plan (C1), counting prices shown after clicking a tab, slider or “see all plans” button.
- Apify, Scrape.do, ScrapingBee, ScraperAPI, Zyte and Storm Proxies were in the August audit but aren’t in this report, which covers residential proxies only. We haven’t re-checked them, so their August results are no longer current grades.
Regrading the old page would have meant putting August and October evidence, and two different rubrics, in one table. The August grades are kept below for the record only.
August 2026 audit grades (retired; pages read 7 August 2026)
| Provider | Grade | Full tiers | Readable | Consistent | Framing | Conditions |
|---|---|---|---|---|---|---|
| Apify | A 5/5 | ✓ | ✓ | ✓ | ✓ | ✓ |
| DataImpulse | A 5/5 | ✓ | ✓ | ✓ | ✓ | ✓ |
| Rayobyte | A 5/5 | ✓ | ✓ | ✓ | ✓ | ✓ |
| Scrape.do | A 5/5 | ✓ | ✓ | ✓ | ✓ | ✓ |
| ScrapingBee | A 5/5 | ✓ | ✓ | ✓ | ✓ | ✓ |
| SOAX | A 5/5 | ✓ | ✓ | ✓ | ✓ | ✓ |
| Webshare | A 5/5 | ✓ | ✓ | ✓ | ✓ | ✓ |
| Bright Data | B 4/5 | ✓ | ✓ | ✓ | ✗ | ✓ |
| Decodo (ex-Smartproxy) | B 4/5 | ✓ | ✓ | ✗ | ✓ | ✓ |
| IPRoyal | B 4/5 | ✗ | ✓ | ✓ | ✓ | ✓ |
| Oxylabs | C 3/5 | ✗ | ✓ | ✓ | ✓ | ✗ |
| Proxy-Seller | C 3/5 | ✗ | ✓ | ✓ | ✓ | ✗ |
| ProxyEmpire | C 3/5 | ✗ | ✓ | ✓ | ✓ | ✗ |
| NodeMaven | D 2/5 | ✗ | ✓ | ✓ | ✗ | ✗ |
| Proxy-Cheap | D 2/5 | ✗ | ✓ | ✓ | ✗ | ✗ |
| ScraperAPI | not scored | — | ✗ | — | — | — |
| Storm Proxies | not scored | — | ✗ | — | — | — |
| Zyte | not scored | — | ✗ | — | — | — |
✓ pass · ✗ fail · — not assessed. Grade = share of assessed criteria passed; providers with fewer than three assessable criteria were not graded.
Methodology
- Providers. The 12 residential proxy providers we track in The Price of a Gigabyte: Bright Data, Oxylabs, Decodo, IPRoyal, SOAX, Webshare, Rayobyte, DataImpulse, NodeMaven, Proxy-Cheap, ProxyEmpire, Proxy-Seller. We looked at rotating residential proxies only.
- Live pages. On 2 October 2026 we loaded each residential pricing page in headless Chrome and saved the HTML, the visible text and a full-page screenshot. Where prices sit behind a tab, slider or button (Proxy-Cheap’s rotating residential tab, NodeMaven’s plan slider, Proxy-Seller’s “See all plans”) we scripted the click and saved what appeared. We saved terms, refund and KYC pages the same way. IPRoyal’s terms and KYC pages and one Bright Data help page redirected headless Chrome to another page, so we fetched those with a plain HTTP request. All 97 quotes on this page were checked automatically against the saved copies.
- History. For eight providers with readable archives we fetched the first Wayback Machine capture of each month, July 2024 to September 2026, of the same pricing page. We counted a discount as shown when the page had a struck-through price or a coupon code, using patterns we checked by hand for each provider. Pages that render prices only with JavaScript (Oxylabs before February 2025, NodeMaven from May 2026) are marked unreadable.
- Scoring. Each criterion is pass/fail; N/A only when the evidence can’t be checked. Criteria met ÷ criteria assessed. No weighting. A pass needs a saved quote; a fail for something missing needs a search of the saved pricing page that finds nothing relevant.
- Neutrality. The same rubric and the same searches were applied to every provider, and every mark can be traced to a saved page in the CSVs. If a provider’s page has changed or we have misread it, we will re-check and note any correction on this page.
Limitations
- A single day. Prices, codes and terms change often. Everything here describes the pages on 2 October 2026.
- Monthly sampling. One capture a month can miss short changes. Some months have no capture; NodeMaven and Proxy-Cheap couldn’t be checked for C7.
- Public pages only. We didn’t create accounts or reach checkout, so we can’t see in-dashboard prices, the tax added at checkout, or how refunds and KYC work in practice.
- What is a “plan”. SOAX sells credit plans with a per-GB rate per plan; Proxy-Seller shows only a “from” price until you click “See all plans”. We counted both as showing per-GB prices.
- Not a quality ranking. The score measures how clearly prices and terms are disclosed, not network quality or value.
Download the data
- Scores (CSV): one row per provider, the result for each criterion and the score.
- Scorecard detail (CSV): one row per provider and criterion, with the result, note, evidence IDs and URLs.
- Evidence (CSV): all 97 quotes with source URL, snapshot file, capture time and HTTP status.
- Absence checks (CSV): the searches behind every “not stated” finding.
- Discount history (CSV): every monthly Wayback capture we used, whether it was readable and what discount it showed.
- Webshare plan prices (CSV): struck price, current price and highest undiscounted price found, per plan.
Free to use with attribution (CC BY 4.0). Quotes remain the property of their publishers.
Cite or embed
Suggested citation: ProxyPicker (2026). “Proxy Pricing Transparency Report: struck-through prices, coupon codes, tax, expiry, refund and KYC terms at 12 residential proxy providers.” Data checked 2 October 2026. https://proxypicker.com/proxy-pricing-transparency-report/
Embed the scorecard:
<a href="https://proxypicker.com/proxy-pricing-transparency-report/"><img src="https://proxypicker.com/wp-content/uploads/2026/10/proxypicker-transparency-2026-10-transparency-scorecard.png" alt="Proxy pricing transparency scorecard, 12 residential proxy providers (ProxyPicker)" width="800" /></a><br />Source: <a href="https://proxypicker.com/proxy-pricing-transparency-report/">ProxyPicker, Proxy Pricing Transparency Report</a> (2 October 2026)Related: The Price of a Gigabyte (residential proxy prices since 2019) · What Your Wi-Fi Is Worth (what bandwidth-sharing apps pay) · Best residential proxies · Scraping Law Hub.
Press & data requests: email [email protected] for interviews, the raw data behind this page, or corrections. See also our contact page.