Meta description: The top web unlocker APIs in 2026 compared: Bright Data, Oxylabs, Zyte, Decodo, ZenRows, Thordata and more, by billing model, success rules and features.
A web unlocker API takes a URL and returns the page, even when the site sits behind Cloudflare, DataDome, Akamai or another anti-bot system. Behind a single endpoint, the vendor picks the proxy, builds a browser fingerprint, solves or avoids CAPTCHAs, renders JavaScript when needed and retries until the request succeeds. You stop maintaining an unblocking stack and pay per successful request or per gigabyte instead.
The products look similar on the surface and differ a lot in the details. Some bill per request, some per gigabyte, and one lets you choose. Some charge only for successful responses, some count 404 pages as a success, and credit multipliers for JavaScript rendering or premium proxies can make the real cost on a protected site ten times the headline rate. Where the proxy IPs come from has also become a buying question after the 2026 takedowns of the IPIDEA and NetNut residential proxy networks.
This guide to the top web unlocker APIs in 2026 compares the providers we would shortlist, based on each vendor’s own site, pricing page and documentation as checked in November 2026. For the ground rules on avoiding blocks before you pay to beat them, see our guide to web scraping best practices.
Web Unlocker APIs At A Glance
Company
Headquarters
Best For
Standout Trait or Pricing Model
Bright Data
Netanya, Israel
Enterprises facing the hardest targets
$1.50 per 1,000 successful requests pay as you go
Oxylabs
Vilnius, Lithuania
Light pages at high volume
Billed per GB, from about $7 down to $3
Zyte
Cork, Ireland
Scrapy teams with mixed-difficulty sites
Priced by site difficulty, per successful response
Decodo
Vilnius, Lithuania
Mid-market teams that want billing choice
Per-GB or per-request plans on the same product
ZenRows
Bilbao, Spain
AI agents and developer tooling
Credit plans from $19 a month; failed requests free
Thordata
Hong Kong
Price-sensitive teams
$1.30 per 1,000 falling to $1.00; success-only billing
ScraperAPI
Not published
Very large crawls
Async queue that retries until success; from $49 a month
Scrape.do
Ankara, Türkiye
Lowest entry cost per request
Hobby plan $29 for 250,000 credits; success-only
ScrapingBee
France
Simple developer experience
Plans from $19 a month; JS rendering on by default
ScrapingAnt
Poland
Prototyping on a free tier
10,000 free credits a month; failed requests free
Best Web Unlocker APIs In 2026
1. Bright Data — The Largest IP Pool Behind A Pay-For-Success Endpoint
Bright Data’s Web Unlocker is the reference product in the category. One endpoint handles proxy selection, fingerprinting, CAPTCHA solving for reCAPTCHA, hCaptcha and Cloudflare Turnstile, automatic retries and JavaScript rendering. It draws on a residential pool of more than 400 million IPs, and parsing, scheduling and webhook delivery sit on top for teams that want structured output rather than raw HTML.
Billing is per request and only for successful ones. Pay as you go is $1.50 per 1,000 requests with unlimited concurrency, and the Scale plan costs $499 a month for 383,000 requests, then $1.30 per 1,000. A free tier includes 5,000 requests a month. Bright Data, founded in 2014, is also the most expensive option here for simple sites, so it earns its place on the targets where cheaper tools fail.
Solvers for reCAPTCHA, hCaptcha and Cloudflare Turnstile
Only successful requests are billed
Free tier of 5,000 requests a month
Best for: enterprises whose targets defeat cheaper unblockers.
2. Oxylabs — Per-GB Unblocking With AI-Driven Fingerprints
Oxylabs Web Unblocker works like a proxy endpoint, so existing scrapers can use it with little change. It generates dynamic browser fingerprints, handles CAPTCHAs, retries automatically and renders JavaScript. Geo-targeting covers 195 countries at country, city or coordinate level, and browser instructions let you click, type and scroll before the page is returned.
It is the main per-GB unlocker on this list. The Starter plan costs $30 a month at about $7.20 per GB, falling to $3 per GB at the top volume tier, which makes it cheap for light HTML pages and costly for heavy ones. Oxylabs, founded in Vilnius in 2015, offers a one-week free trial. Ask how blocked responses are billed, since third-party testing and the product page describe it differently.
Proxy-style endpoint that drops into existing code
Country, city and coordinate targeting in 195 countries
One-week free trial
Best for: high-volume scraping of lightweight pages.
3. Zyte — Pricing That Follows How Hard Each Site Is
Zyte, founded in Cork in 2010 by the team behind the open-source Scrapy framework, sells Zyte API with automatic ban handling built in. It adjusts proxies, request patterns and fingerprints per site and returns either a plain HTTP response or a page rendered in a pre-warmed headless browser. Native Scrapy integration makes it the natural choice for Python teams already using that framework.
Its pricing is the most granular here. Each site is placed in one of five difficulty tiers, and you pay per successful response. Pay-as-you-go HTTP requests range from $0.13 to $1.27 per 1,000, browser-rendered pages from $1.01 to $16.08 per 1,000, and committed plans go far lower. Easy sites end up very cheap, and the hardest ones cost what they cost everywhere. A $5 free credit lets you see which tier your targets fall into.
Five price tiers set by site difficulty
Billed per successful response
Native Scrapy integration
Best for: Scrapy users scraping a mix of easy and hard sites.
Decodo, the Vilnius company that rebranded from Smartproxy in April 2025, sells Site Unblocker as a proxy endpoint with automatic anti-bot handling and an optional JavaScript rendering switch. It covers more than 195 locations, supports sessions of up to 10 minutes, and can return Markdown or screenshots as well as HTML. The dashboard shows failed requests and success rates, which helps when you are testing new targets.
Its distinctive feature is billing choice: the same product is sold on either a per-GB or a per-request plan, so you can match the billing unit to your page weight. Decodo’s documentation notes that requests sent with forced headers or cookies are billed even if they fail, so check how your scraper sends requests before you estimate costs.
Per-GB or per-request plans
Markdown and screenshot output
Success-rate reporting in the dashboard
Best for: mid-market teams that want to pick the billing unit that suits their pages.
ZenRows, based in Bilbao, now sells its unblocking API as Fetch, the successor to its Universal Scraper API. It handles anti-bot systems, JavaScript rendering, geo-targeting and sessions, and an adaptive stealth mode chooses the right setup for protected sites automatically. An MCP server, CLI and SDKs make it easy to hand the API to an AI agent.
Billing uses credits: one for a basic request, five with JavaScript, ten with premium residential IPs and 25 with both. Failed and retried requests are never charged, though 404 and 410 responses count as successes. The free tier has 5,000 credits, the Build plan costs $19 a month, and Enterprise plans carry SOC 2 and ISO 27001.
Adaptive stealth mode for protected sites
MCP server, CLI and SDKs
Failed and retried requests never charged
Best for: developers wiring web access into AI agents.
6. Thordata — A Low Per-Request Rate With Success-Only Billing
Thordata’s Web Unlocker follows the standard model: send a URL and get the page back as HTML or a PNG screenshot, with fingerprint simulation, CAPTCHA solving, retries and IP rotation handled for you. JavaScript rendering is included, and code samples cover seven programming languages, so getting a first request working is quick.
Pricing is the reason to shortlist it. Rates start at $1.30 per 1,000 responses at the 10,000 tier and fall to $1.00 per 1,000 at 200,000, and only successful requests are billed. A seven-day free trial is available. The unlocker sits alongside Thordata’s scraping browser, SERP API and scraper APIs on the same account, which helps teams that want to compare approaches without opening several vendor accounts.
From $1.30 down to $1.00 per 1,000 requests
Success-only billing and a seven-day trial
HTML or screenshot output
Best for: cost-conscious teams that want a dedicated unlocker at a low per-request rate.
7. ScraperAPI — An Async Queue That Retries Until It Succeeds
ScraperAPI, founded in 2018, wraps proxy rotation, CAPTCHA handling and optional JavaScript rendering into one API call, drawing on more than 40 million proxies in over 50 countries. Its Async API accepts large batches of URLs and keeps retrying in the background until each succeeds, then delivers the results, which removes most of the retry logic from your own code.
The Hobby plan costs $49 a month for 100,000 credits. A normal request costs one credit, rendering or premium proxies cost ten, and the toughest ultra-premium setting with rendering costs 75. Only 200 and 404 responses are charged, and geotargeting is free, though the two lowest plans are limited to US and EU locations. Structured endpoints for Amazon, Google and Walmart return JSON rather than HTML.
Async API for very large URL batches
Only 200 and 404 responses billed
Structured JSON endpoints for major sites
Best for: large crawls where you want the vendor to own retries.
8. Scrape.do — The Clearest Success-Only Billing Policy
Scrape.do, founded in Ankara in 2020 and now moving its headquarters to the US, bypasses web application firewalls and anti-bot systems with dynamic TLS fingerprinting, CAPTCHA handling and a mix of residential, mobile and datacenter proxies. Geotargeting covers more than 160 countries, JavaScript rendering is on every plan, and the Advanced plan adds custom bypass strategies for difficult sites.
It has the lowest entry price per request on this list: the Hobby plan costs $29 a month for 250,000 credits, about $0.11 per 1,000 basic requests. Billing is strictly success-only. Timeouts, blocks and errors are free, and support will refund credits for a 200 response that returned unusable content. A free tier includes 1,000 credits.
Lowest entry cost per request here
Blocks, timeouts and errors never billed
JS rendering on every plan
Best for: small teams that want predictable success-only costs.
9. ScrapingBee — The Simplest Developer Experience
ScrapingBee, founded in France in 2019 and now part of the Oxylabs group while running as its own brand, is known for an API that is easy to get started with. Headless JavaScript rendering is on by default, and premium and stealth proxy tiers handle protected sites. Extraction rules and screenshots are built in, and dedicated APIs cover Google, Amazon and ChatGPT.
The Hobby plan costs $19 a month for 75,000 credits. Stealth proxy requests cost 75 credits each, so protected sites burn through a plan faster than the headline suggests. Credits are consumed by successful requests, and Auto Mode costs nothing if every configuration fails. Geotargeting on premium proxies covers more than 200 countries.
JS rendering on by default
Extraction rules and screenshots built in
Plans from $19 a month
Best for: developers who want to get to clean data with the least setup.
ScrapingAnt, operated by DATAANT and headquartered in Poland, runs every request in a real, current-version headless Chrome on rotating datacenter or residential proxies, with a pool of more than three million rotating proxies. An MCP server connects it to Claude Code and Cursor, so developers can give an AI coding agent working web access in minutes.
The free tier includes 10,000 credits a month, the most generous on this list, and the Enthusiast plan costs $19 a month for 100,000 credits. Credit costs range from one for a plain request to 125 for a browser request with residential IPs. Failed requests cost nothing.
10,000 free credits every month
Real headless Chrome on every request
MCP server for AI coding agents
Best for: developers prototyping on a free tier before committing.
How To Choose A Web Unlocker API
Headline prices say little about what a web unlocker will cost on your targets. These questions get to the real number.
How Are You Billed, And What Counts As Success?
Per-request billing suits heavy pages, and per-GB billing suits light ones. Then check the success rules: whether blocks and 4xx responses are free, whether a 404 counts as a success, and whether custom headers or cookies void the guarantee. Two vendors with the same headline rate can differ several times over once these rules apply.
What Will Your Own Targets Actually Cost?
Credit multipliers for JavaScript rendering and premium proxies, and difficulty tiers such as Zyte’s, mean protected sites can cost ten to 25 times the base rate. Use free credits to run the same set of your real URLs through two or three vendors and compare cost per successful page.
Do You Need Specific Locations?
Check country counts, whether city or coordinate targeting is available, and whether low-tier plans restrict locations. Prices, stock and search results often change by location, so geo control matters for e-commerce and SEO work.
Where Do The Vendor’s IPs Come From?
After Google and law enforcement disrupted the IPIDEA and NetNut residential proxy networks in 2026, ask every vendor how its residential IPs are sourced and how consent is obtained, and ask for certifications such as SOC 2 or ISO 27001. Our explainer on proxy types covers the background, and our overview of whether web scraping is legal covers the collection side.
Conclusion
Bright Data remains the benchmark for the hardest targets, Oxylabs is the strongest per-GB option, and Zyte’s difficulty pricing rewards teams with many easy sites. Decodo offers billing flexibility, ZenRows leads on agent tooling, and Thordata, Scrape.do, ScrapingBee, ScraperAPI and ScrapingAnt give cost-conscious teams several solid ways in.
Shortlist two or three, test them on your own URLs, and compare cost per successful page rather than list price. Collect only what you need and respect the terms and rights of the sites you access.
