Our Review Process
Last updated:
Every score on ProxyGlance comes from the same documented process. This page explains what we test, how each criterion is weighted, and how we keep scores current. If a vendor's score looks wrong to you, this page should tell you why it is what it is — and how to challenge it.
How scoring works
- Scores are per product, not per company. Many vendors sell proxies, a scraping API and sometimes an antidetect browser. Each product is scored separately against criteria built for its category. A company can score 9.1 for proxies and 7.4 for its scraping API, and we never blend those into a single number.
- Each criterion is scored from 0 to 10, in half-point steps, against the rubric below.
- The overall score is a weighted average of the criteria, rounded to one decimal place. The weights are fixed per category and are the same for every vendor.
- Scores are comparable within a category only. An 8.5 proxy provider and an 8.5 antidetect browser were judged on different criteria.
What the numbers mean
| Score | Meaning |
|---|---|
| 9.0 – 10 | Excellent. Leads its category on most criteria with no significant weaknesses. |
| 8.0 – 8.9 | Very good. A strong choice for most use cases, with minor trade-offs. |
| 7.0 – 7.9 | Good. Solid for specific use cases; clear weaknesses elsewhere. |
| 6.0 – 6.9 | Fair. Works, but most buyers have better options. |
| Below 6.0 | Not recommended in its current form. |
Proxy providers
| Criterion | Weight | What we measure |
|---|---|---|
| Performance | 25% | Success rate against our test targets, response time, connection stability and session persistence. |
| Pool & Coverage | 20% | Pool size, IP uniqueness across requests, country count and geo-targeting accuracy. |
| Pricing & Value | 20% | Effective price per GB or per IP at common plan sizes, minimum commitment, trial terms and refund policy. |
| Features | 10% | Targeting options (country, city, ASN, ZIP), rotation control, protocols, usage stats and API access. |
| Ease of Use | 10% | Sign-up and setup time, dashboard clarity, documentation and code examples. |
| Support | 5% | Available channels, response time and the quality of answers to our test questions. |
| Trust & Compliance | 10% | IP sourcing transparency, KYC and abuse policies, certifications and company track record. |
How we test. We buy or request a standard plan for each proxy type a provider sells and send requests through it to a fixed set of targets that range from unprotected pages to sites with active bot protection. We record success rate, response time and connection errors, check the reported location of the exit IP against the location we asked for, and measure how often the same IP repeats across a sample of requests. Pricing is calculated at common plan sizes, not the cheapest headline rate, because per-GB prices often depend on a large monthly commitment.
Many providers sell ISP and static residential proxies as the same product. We test and list them under both types when a provider does, and explain the overlap in our ISP and static residential guides.
Scraping API providers
| Criterion | Weight | What we measure |
|---|---|---|
| Performance | 30% | Success rate on protected and unprotected targets, response time and consistency across repeated runs. |
| Target Coverage | 15% | Supported sites and dedicated endpoints (SERP, e-commerce, social), geo-targeting and JavaScript rendering. |
| Pricing & Value | 20% | Effective cost per 1,000 successful requests, credit multipliers for rendering or premium targets, and trial terms. |
| Features | 15% | Structured (parsed) output, browser actions, async/batch jobs, webhooks and SDKs. |
| Ease of Use | 10% | Time to first successful request, API design, documentation and dashboard. |
| Support | 5% | Available channels, response time and the quality of technical answers. |
| Trust & Compliance | 5% | Data handling policies, acceptable-use rules, certifications and company track record. |
How we test. We run the same batch of requests against each API — general pages, search engine results, e-commerce product pages and JavaScript-heavy sites — and count a request as successful only when it returns the content we asked for, not just an HTTP 200. Where an API charges extra credits for JavaScript rendering, premium proxies or protected targets, we include those multipliers in the effective cost per 1,000 successful requests.
Antidetect browsers
| Criterion | Weight | What we measure |
|---|---|---|
| Fingerprint Protection | 30% | Results on public fingerprint and bot-detection checks, and consistency of the generated fingerprints. |
| Platform Support | 10% | Desktop operating systems, browser engines and how closely engine versions track upstream releases. |
| Pricing & Value | 20% | Price per profile and per team seat, free tier limits and how plans scale. |
| Features | 15% | Profile management, proxy integration, cookie import/export, automation API and team permissions. |
| Ease of Use | 10% | Install and onboarding, creating and launching profiles, interface clarity and documentation. |
| Support | 5% | Available channels, response time and the quality of answers to our test questions. |
| Trust & Compliance | 10% | Security practices, data storage (local vs cloud), company transparency and track record. |
How we test. We install each browser on a clean machine, create profiles with default and custom settings, and run them through public fingerprinting and bot-detection checks. We look for inconsistencies between reported values (for example, a time zone that does not match the proxy location) and for how closely the browser engine tracks the current Chrome or Firefox release. We also test the automation API, proxy integration and team features where they exist.
Keeping data current
Prices, pool sizes and features change often. Every vendor page shows a last reviewed date, and we:
- Re-check pricing and plan details when a vendor announces changes, and at least every six months.
- Re-run performance tests when a product changes significantly or when readers report problems we can reproduce.
- Update the score and the last reviewed date together, so you can see when a number was last confirmed.
All facts on the site — prices, pool sizes, scores — come from one record per vendor. When we update a vendor's price, it changes on its review, every directory and every ranking at the same time.
Independence
We earn commissions through some of the links on this site. That does not change how products are tested or scored:
- Vendors cannot pay for a review, a higher score or a better ranking position.
- Products without an affiliate programme are reviewed and ranked with the same criteria as those with one.
- Vendors may send corrections to facts such as prices or features. We verify them before changing anything, and they never see a review before it is published.
Read our full affiliate disclosure for how the site is funded.
Corrections
If you think a fact or score is wrong, email hello@proxyglance.com with the page and what you believe is incorrect. We review every correction request, and when we change a score we update the last reviewed date on that page.