Zyte
The world's preeminent rotating proxy network, Zyte's Smart Proxy Manager ensures your web data is delivered quickly and reliably.
Summary
Zyte Smart Proxy Manager provides a single API for web scraping that handles proxy rotation and anti-bot bypass. Developers can perform headless browser rendering, execute custom actions like scrolling, and capture screenshots via JSON payloads. The tool supports AI-powered data extraction for products and articles. It offers geotargeting across 249 countries and optional residential IP routing.
Zyte’s Products
Zyte User Reviews
Zyte offers a comprehensive range of web scraping and data collection products, from proxies to data sets and data APIs.
User Review Summary
Reviewers frequently praise the technical support for being responsive and professional, though some users report slow or unhelpful communication. Many customers value the reliability of the data extraction and proxy services for bypassing blocks. However, there are consistent criticisms regarding billing practices, including unexpected charges and the lack of functional spending limits. While some users find the performance scalable, others disagree, citing issues with latency and throttled request speeds after upgrading to paid plans.
The best web scraping service I’ve used
The service works well even with websites that have complex and custom anti-bot protection.
What I especially like is the pricing model: Zyte automatically determines whether a request was successful and only charges for successful requests. This is very convenient and makes costs much more predictable when scraping difficult websites.
In a few particularly challenging cases, I contacted support, and they helped me choose the right configuration parameters and get everything working reliably.
Overall, Zyte is a very convenient and reliable service, especially if you need to work with websites that have strong anti-bot protection.
Since we started working with Zyte
Since we started working with Zyte, we have consistently experienced a level of support that truly stands out and surpasses that of their peers. While many companies can respond to support tickets, few can meet the level of responsiveness and care that clients need when facing critical issues.
The Zyte team demonstrates a genuine sense of urgency and empathy. They quickly step in to help, keep us informed throughout the resolution process with regular progress updates, and, most importantly, deliver the fixes needed to resolve issues efficiently.
I ran into a few issues at first
I ran into a few issues at first, but the Zyte team was professional and very helpful throughout. They communicated clearly and helped me get everything sorted quickly. I’m happy with the support I received and the final outcome. Would recommend.
Zyte API
Zyte API is a web scraping and data extraction tool reached via an HTTP endpoint or proxy port. It handles proxy rotation, headless browser management, and anti-bot bypass, while providing AI-powered extraction for products and other structured data.
Ideal Use Cases
Zyte API is a scraping API accessed through a single HTTP endpoint or a proxy gateway. It is used for automated data extraction and browser rendering.
Ideal Use Cases
- E-commerce & price monitoring: Built-in product and productList automated extraction fields.
- JavaScript-heavy sites & SPAs: Headless browser rendering enabled via browserHtml: true.
- Screenshots & visual monitoring: Captured after JS execution via screenshot: true.
- Real estate listings: Supports automated extraction and browser actions for navigation.
Poor Use Cases
- Large-scale crawling: Per-request cost multipliers for JS rendering may exceed budget.
- Real-time, low-latency lookups: Browser rendering and anti-bot checks add significant latency.
- Mobile apps & mobile-only content: Default device emulation is desktop.
- Sneaker & ticket purchasing bots: Terms of service usually prohibit checkout automation.
Performance
Assuming average page size is 250kb
| Domain | Feature | Success Rate | Avg. Success Latency | API Credits | Provider Plan | CPM |
|---|---|---|---|---|---|---|
| aliexpress.com | - | 100% | 14.1s | 2 credits | $100 Per Month ($100/month) | $100 |
| amazon.com | - | 94% | 13.3s | 2 credits | $100 Per Month ($100/month) | $100 |
| apple.com | - | 100% | 3.1s | 2 credits | $100 Per Month ($100/month) | $100 |
| etsy.com | - | 100% | 15.9s | 3 credits | $100 Per Month ($100/month) | $100 |
| target.com | - | 100% | 11.8s | 3 credits | $100 Per Month ($100/month) | $100 |
| walmart.com | - | 99% | 13.8s | 4 credits | $100 Per Month ($100/month) | $100 |
| zalando.co.uk | - | 100% | 10.9s | 1 credit | $100 Per Month ($100/month) | $100 |
Fingerprint Resistance Tests
Zyte API ranked #3 of 9 proxy APIs on our 15 fingerprint tests (analysis of Nov 5, 2026). Every proxy API loaded deviceandbrowserinfo.com 5 times from the US, the UK and Germany through the ScrapeOps Proxy API (JavaScript rendering on, residential proxies, 15 s wait), so its own headless browser was tested, and each test was graded against the same fixed rubric.

How Zyte API compares with the other proxy APIs
Overall score on the 15 fingerprint tests, out of 100. Higher is better.
In our benchmark analysis, Zyte API ranked #3 out of 9 proxy APIs with an overall score of 68. It demonstrated strong performance in core automation masking and platform consistency but struggled with certain hardware realism and header configuration tests.
| Test | Status |
|---|---|
| Automation Signals | Pass |
| TLS / JA3 Realism | N/A |
| Platform Consistency | Pass |
| Device Type Coherence | Pass |
| Graphics Fingerprints | Fail |
| Hardware Realism | Fail |
| Client Hints Coherence | Pass |
| Fingerprint Entropy | N/A |
| Test | Status |
|---|---|
| Header Realism | Fail |
| Peripherals Presence | Pass |
| Fonts & Plugins | Pass |
| Resolution & DPR | Pass |
| Timezone vs IP Geo | Fail |
| Viewport/Geometry | Pass |
| Language/Locale vs IP | Warn |
- ✅ Where Zyte API performed well: Successfully masked all automation signals (Webdriver/CDP); maintained consistent platform reporting across HTTP and JS layers; and provided realistic screen resolutions and peripheral counts.
- ❌ Where Zyte API fell short: Failed to consistently mask software-based graphics drivers; exhibited a critical missing header in some sessions; and showed misalignments in geo-specific language and timezone settings.
Headers and Device Fingerprints
Zyte API provided a generally solid foundation for stealth, successfully hiding the most common signs of automated control. In all tested sessions, indicators like navigator.webdriver remained false, and the browser environment appeared to be a genuine Windows desktop instance.
However, the environment's realism was undermined by the exposure of software-rendering drivers in a subset of sessions, which serves as a signal for server-side execution. While many sessions used real consumer GPUs, the inconsistency across the pool creates a detectable pattern.
Geographic alignment also proved to be a challenge. While the platform and device types were coherent, the browser often defaulted to US-centric language and timezone settings regardless of the proxy exit node's actual location.
The header set was modern and typically well-formed, but the occasional omission of standard encoding headers represented a lapse in protocol realism that advanced anti-bot systems may flag.
Good
In our tests, Zyte API showed sophisticated handling of many high-weight fingerprinting categories.
-
✅ Clean Automation Profile: No sessions exposed
webdriver: trueor CDP automation flags. The web worker environment was also fully consistent with the main frame, ensuring no leaks in the JavaScript execution context. -
✅ Consistent OS Triplet: The platform was reported consistently as "Windows" across the User-Agent,
navigator.platform(Win32), and Client Hints, avoiding the cross-layer contradictions common in lower-tier providers. -
✅ Realistic Viewport & Resolution: Screen sizes were modern and varied (e.g.,
1920x1080,2560x1440), with inner window dimensions that correctly accounted for browser toolbars and taskbars.
Screen: 1920x1080 Available: 1920x1040 Inner: 929x925 // Properly nested geometry indicates a realistic browser UI.
- ✅ Peripheral Presence: Unlike many bots that report zero media devices, Zyte API sessions reported microphones, speakers, and webcams, enhancing the profile's authenticity.
Bad
Despite its strengths, Zyte API exhibited several consistency issues that lowered its overall realism score.
❌ Software Graphics Rendering
In some sessions, the browser exposed the Microsoft Basic Render Driver. This is a software-based renderer typically found on servers or virtual machines, rather than the hardware GPUs used by actual consumer devices.
- GPU Renderer:
ANGLE (Microsoft, Microsoft Basic Render Driver (0x0000008C) Direct3D11...)was observed in roughly 13% of sessions. - Vendor: The vendor was reported as
Google Inc. (Microsoft), a combination highly associated with headless environments.
❌ Header Realism Issues
While most sessions were realistic, a failure occurred where a session sent an empty encoding header, which is an impossible configuration for the claimed modern Chrome version.
- Accept-Encoding: One session from the US sent an empty
Accept-Encodingheader. - Impact: Legitimate Chrome browsers (version 156+) always send encoding tokens like
brorzstd.
❌ Timezone & Language Misalignment
Zyte API struggled to align the browser's internal locale and clock settings with the proxy's geographic location.
- Timezone: Sessions in Germany and the UK frequently defaulted to
America/New_YorkorAmerica/Chicago. - Language: A session in Germany incorrectly reported
es-419(Spanish - Latin America).
Target Geo: DE (Germany) Expected Timezone: Europe/Berlin Observed Timezone: America/New_York Observed Language: es-419
❌ Hardware Realism Contradictions
The presence of software rendering drivers mentioned in the graphics section also triggered failures in hardware realism, as real consumer hardware should be paired with physical GPU drivers.
- Contradiction: High-performance hardware (20 cores, 16GB RAM) was paired with the
Microsoft Basic Render Driversoftware renderer.
Verdict: ⚠️ Mixed
During our tests, Zyte API provided a robust but occasionally inconsistent stealth environment.
It excelled at the "basics" of bot detection—successfully hiding automation flags and maintaining OS consistency. However, the intermittent exposure of software renderers and the failure to correctly localize timezone and language settings for non-US traffic represent significant detection risks for high-security targets.
✅ What it gets right
- Full suppression of
navigator.webdriverand CDP flags. - Excellent platform consistency between HTTP and JS layers.
- Realistic screen geometry and peripheral device counts.
❌ What holds it back
- Exposure of software-rendering drivers (
Microsoft Basic Render Driver). - Inconsistent localization (US timezones/languages for European IPs).
- Occasional missing headers in the network layer.
Bottom line Zyte API is a capable contender in the proxy API space, performing well for general-purpose scraping. Users should be aware that its environment occasionally leaks "server-like" traits, making it most effective for targets that do not heavily rely on WebGL or strict geo-fencing checks.
Test details
Click any row to see what the test observed and why it passed or failed.
| Test | Result | Summary | Weight | |
|---|---|---|---|---|
| Automation Signals | Pass | All automation flags were false in all 15 sessions, with no HeadlessChrome User-Agent and consistent worker values, so the browser passed. | 10 | ▸ |
| TLS / JA3 Realism | N/A | TLS / JA3 fingerprints are not collected in this benchmark. | 8 | ▸ |
| Platform Consistency | Pass | All 15 sessions reported Windows consistently across the User-Agent, navigator.platform (Win32), userAgentData (Windows) and web worker, with no missing or non-standard values. | 6 | ▸ |
| Device Type Coherence | Pass | All 15 sessions were consistent Windows desktop profiles (Is mobile false, ?0, widths 1280-2560); only 4/15 reported touch points, below the 50% warn threshold. | 5 | ▸ |
| Graphics Fingerprints | Fail | Two of 15 sessions (1 and 12) exposed the Microsoft Basic Render Driver, a software renderer, so the test fails under decision row 2. | 5 | ▸ |
| Hardware Realism | Fail | Two of 15 sessions reported the software renderer 'Microsoft Basic Render Driver', so the test failed under C3. | 5 | ▸ |
| Client Hints Coherence | Pass | Client Hints matched the User-Agent in all 15 sessions: Windows platform, ?0 mobile flag and matching Chrome major versions. | 3 | ▸ |
| Fingerprint Entropy | N/A | Not scored: the same device settings in every session are normal. | 3 | ▸ |
| Header Realism | Fail | Fail: Accept-Encoding was empty in 1 of 15 sessions (C1), although all User-Agents were current Chrome 156-158 with no bot tokens. | 3 | ▸ |
| Peripherals Presence | Pass | All 15 sessions reported valid numeric peripheral counts of at most 1, and only 1 session had neither a microphone nor a webcam, so C3 did not trigger in every session. | 3 | ▸ |
| Fonts & Plugins | Pass | All 15 Win32 sessions had Windows-consistent fonts and a PDF plugin list, and only 4 of 15 had short font lists, so the test passed. | 2 | ▸ |
| Resolution & DPR | Pass | All 15 sessions reported common modern screen sizes with normal aspect ratios and inner windows that fit within the screen. | 2 | ▸ |
| Timezone vs IP Geo | Fail | 2 | ▸ | |
| Viewport/Geometry | Pass | Screen, available and inner resolutions were present and nested correctly in all 15 sessions, with taskbar and browser toolbar offsets, apart from one session without a taskbar. | 2 | ▸ |
| Language/Locale vs IP | Warning | One DE session used es-419 (wrong language) and 9 of 15 sessions had only an acceptable en-US for DE/UK, giving a warn. | 1.5 | ▸ |
Features
| Feature | API Credit | Description |
|---|---|---|
RenderingbrowserHtml | - | Loads the page in a browser, returns rendered HTML. |
ResidentialipType | - | Forces the request out of a residential IP. |
Rendering + ResidentialipType | - | Browser rendering sent from a residential IP. |
Countrygeolocation | - | Overrides the exit country of the request. |
Browser Waitactions | - | Wait for a selector or a fixed time before output. |
Geotargeting
Geotargeting is managed via the geolocation parameter using ISO 3166-1 alpha-2 codes. It supports 249 countries. While the API selects optimal locations by default, manual overrides for extended geolocations incur additional costs.
| Standard Countries No extra cost | |
|---|---|
| Australia | Italy |
| Belgium | Japan |
| Brazil | South Korea |
| Canada | Mexico |
| China | Netherlands |
| Germany | Poland |
| Spain | Russia |
| France | Turkey |
| United Kingdom | United States |
| India | South Africa |
| … and many more countries | |
Zyte API Plans
Every website belongs to one of 5 tiers, and the tier sets the price. Committing to a monthly minimum lowers every rate in the table.
| Website tier | Pay as you goper 1,000 requests | $100per 1,000 requests | $200per 1,000 requests | $500per 1,000 requests |
|---|---|---|---|---|
| 1Simpleikea.comzara.comnike.com | $0.13 | $0.10 | $0.08 | $0.06 |
| 2Easyamazon.combooking.comairbnb.com | $0.23 | $0.17 | $0.14 | $0.11 |
| 3Moderatetarget.comlowes.comtemu.com | $0.44 | $0.33 | $0.26 | $0.21 |
| 4Complexhomedepot.commacys.comsephora.com | $0.70 | $0.53 | $0.42 | $0.34 |
| 5Advancedcostco.comnordstrom.comexpedia.com | $1.27 | $0.95 | $0.76 | $0.61 |
| Enterprise | Further discounts based on volume usage.Contact sales | |||
- Sites Zyte has too little traffic data on are billed at a blended rate until a tier is assigned.
- A few sites are priced above tier 5; Zyte quotes those individually.
Prices checked Nov 4, 2026.
Integrating Zyte API
Developers connect via a single HTTP POST endpoint or a proxy gateway at api.zyte.com:8011. Authentication is performed using Basic Auth with the API key as the username.
curl \
--user YOUR_ZYTE_API_KEY: \
--header 'Content-Type: application/json' \
--data '{"url": "https://toscrape.com", "httpResponseBody": true}' \
--compressed \
https://api.zyte.com/v1/extractThe proxy gateway uses the API key as the username for authentication. Parameters such as browser rendering are passed through custom HTTP headers like Zyte-Browser-Html.
curl \
--proxy api.zyte.com:8011 \
--proxy-user YOUR_ZYTE_API_KEY: \
--compressed \
https://toscrape.comimport requestsfrom base64 import b64decodeapi_response = requests.post("https://api.zyte.com/v1/extract",auth=("YOUR_ZYTE_API_KEY", ""),json={"url": "https://toscrape.com","httpResponseBody": True,},)http_response_body = b64decode(api_response.json()["httpResponseBody"])print(http_response_body.decode())
Examples