Web scraper reviews for developers who pay the bills: proxies by the GB, scraping APIs by the credit, AI-agent
search by the call. Picks for each use case, with the facts behind them. Every performance number names its
source and test date.
Our top five to start with. Each one links to the facts behind it.
1Simplest flat bill that never expiresPacketStream: One price for everything: $1.00/GB for rotating and sticky traffic, with no tiers and no feature multipliers.
2Best-tested network on pay-as-you-goDecodo: 99.92% success, 0.53 s average and 2,853,543 unique IPs in Proxyway’s April 2026 test.
3Flat, predictable price with rendering includedBright Data Web Unlocker: One flat $1.50 per 1,000 whatever the target needs; pay only for success; 5,000 free requests a month.
4Most complete search + fetch toolkitScrapingBee API + remote MCP: Google search (eight result types) and a sub-second Fast Search, plus fetch with JavaScript rendering, premium proxies, screenshots and extraction: 18 MCP tools.
5Lowest JavaScript-rendering rate on a subscriptionOxylabs Web API: $1.22 per 1,000 rendered pages on larger plans, with premium routing included.
Five guides, one per job
Each guide opens with picks for specific use cases within that job, each with the facts behind it. Use the
chips to show one guide at a time, or scroll through all five.
Filter:
01 / PROXIES
Proxies for web scraping
If you run Scrapy, Playwright or Crawlee yourself, residential bandwidth is the cost that grows with every page. PacketStream has the simplest bill: $1/GB for rotating and sticky traffic alike, with no subscription and a balance that never expires. Decodo has the best recent test results on pay-as-you-go, IPRoyal the longest sticky sessions and Bright Data the finest geo-targeting.
$4/GB pay-as-you-go; plans from $3.75/GB (3 GB) to $2.75/GB (100 GB) (as of October 2026)
99.92% success, 0.53 s average and 2,853,543 unique IPs in Proxyway’s April 2026 test.
Pay-as-you-go available; no mandatory KYC for standard use; 24/7 support.
Caveats: Four times the entry price of Evomi Core or PacketStream.
Runner-up:Oxylabs. If a subscription is fine: 99.93% success, 0.60 s and the largest measured pool (3.2M+ unique IPs) in Proxyway’s testing, from $30/mo.
Six scraping vendors run MCP servers that let Claude Code, Cursor or ChatGPT search the web and read pages. PacketStream puts Google search, Markdown fetch and a hosted MCP server on one never-expiring prepaid balance, at $2.50 per 1,000 successful calls with failed calls free. Bright Data’s SERP API has the lowest pay-as-you-go Google search price ($1.50 per 1,000, successes only); ScrapingBee is the most complete subscription toolkit; ZenRows and Firecrawl have free monthly allowances.
Best for
Google search, fetch and MCP on one prepaid balance
$2.50 per 1,000 successful calls; every failed call free (as of October 2026)
$0.0025 per successful search or fetch, and failed calls cost nothing.
Google web, news, shopping, places, autocomplete, Scholar and AI Mode results as JSON or Markdown; Fetch returns a page’s main content as Markdown with its title, description and final URL.
Hobby $19/mo (75K credits): Google search 10 credits, plain fetch 1, JS fetch 5 (as of October 2026)
Google search (eight result types) and a sub-second Fast Search, plus fetch with JavaScript rendering, premium proxies, screenshots and extraction: 18 MCP tools.
On a subscription, 25,000 searches fit the $49 Freelance plan ($1.96 per 1,000), alongside JS-rendering fetches.
Caveats: Subscription only; unused credits expire, so light use still costs $19 a month.
PacketStream’s Fetch API turns public pages into clean Markdown for $2.50 per 1,000 fetches that return content, with every error free. When the target runs Cloudflare, Akamai, DataDome or Kasada, you pay per successful page instead of fighting blocks yourself: Zyte had the highest success rate in the latest independent benchmark (October 2025), and Bright Data’s Web Unlocker has a flat $1.50 per 1,000 with rendering included, backed by a 2024 result.
For people who need data, not a codebase. If you run Apify Actors, PacketStream’s $1/GB residential proxies can be entered as your own proxies in any Actor whose input offers proxy configuration, against the $8/GB list price of Apify’s own residential proxy. Apify brings thousands of ready-made scrapers and an API; Octoparse is the classic point-and-click desktop app; ParseHub still works but shows little development since 2023.
$1.00/GB flat, prepaid from $50, balance never expires; Apify’s own residential proxy lists at $8/GB (as of October 2026)
$1.00/GB flat against $8/GB ($7 on higher plans) for Apify’s own residential proxy, with no subscription and a balance that never expires.
Apify accepts your own proxies: in an Actor’s Input tab, open “Proxy and browser configuration” and enter your proxy URLs. It works in Actors whose input offers proxy configuration.
Free (10 tasks, local runs); Standard from $69/mo (check the billing period); Professional $249/mo (as of October 2026)
A no-code desktop app with a visual builder; tasks run locally or in Octoparse’s cloud.
Residential proxies ($3/GB) and CAPTCHA solving ($1–1.50 per 1,000) available as add-ons; 5-day money-back.
Caveats: Monthly, quarterly and annual billing are offered (annual saves 16%), and the page doesn’t make clear which period the $69 assumes.
Runner-up:ParseHub. A long-running visual builder with a free plan of 200 pages per run; paid plans start at $189/mo and development has been quiet since 2023.
Free to use, paid for in servers, proxies and your time. PacketStream supplies the proxies at $1/GB flat, with setup guides for Scrapy, Playwright and Puppeteer. Crawlee bundles browser crawling with proxy rotation and sessions, Scrapy is the fastest way to crawl HTML at scale in Python, and Playwright (or the lighter Puppeteer) scripts a real browser.
Best for
Residential proxies with setup guides for Scrapy, Playwright and Puppeteer
Chromium, Firefox and WebKit from several languages; 1.64.1 released 14 Oct 2026.
Proxy per browser or per context (HTTP(S) or SOCKS5) and route.abort() to block heavy assets.
Caveats: Not a crawler: queueing, retries and storage are yours to build.
Runner-up:Puppeteer. Lighter, Chrome-focused automation for Node.js (25.14.0, 14 Oct 2026); the popular stealth plugin has had no release since March 2023.
Each guide names picks for concrete use cases ("simplest flat bill that never expires", "Google search, fetch and
MCP on one prepaid balance") and backs each one with checkable facts: the vendor's own price and terms,
documented features, and dated third-party test results, linked from every pick.
Prices were checked on each vendor's pricing page on 17 October 2026. Performance figures come from Proxyway and
ScrapeOps, which have commercial interests of their own; we cite them with their test dates and say so. The
methodology page sets out the evidence behind every pick and where it comes from.
What scraping really costs
Running your own scraper, bandwidth is the line that grows with every page: page size × pages × price per GB. A
1,000-page HTML-only crawl (about 100 KB a page) costs $0.10 at $1/GB, $0.40 at the $4/GB market median and $0.80 at
$8/GB. Switch to full headless renders (about 2 MB) and the same 1,000 pages cost $2 to $16. Blocking images, fonts
and CSS is the cheapest saving available.
Managed APIs flip the model: you pay per successful page, and the price depends on what the page needs. Plain HTML can
cost 13–28 cents per 1,000; JavaScript plus premium proxies on a hard site can cost $5–$37 per 1,000 on entry plans,
or a flat $1.50 at Bright Data's unblocker. The web scraping cost
calculator compares the two for your volume, and the build vs buy guide
covers servers, CAPTCHAs and maintenance time.
FAQ
What is the best web scraping tool?
It depends on the job: proxy bandwidth for your own scraper, a scraping API, search and fetch tools for an AI agent, a no-code app or an open-source framework. Our first pick for most readers is PacketStream: residential proxies at $1/GB flat for rotating and sticky traffic, a prepaid balance that never expires, and Google Search, Markdown Fetch and MCP APIs ($2.50 per 1,000 successful calls) on the same balance. Each guide then gives picks for specific use cases, such as Decodo for a pay-as-you-go proxy network with strong recent test results, Bright Data’s Web Unlocker for a flat $1.50 per 1,000 with rendering included, and ScrapingBee for a complete search and fetch toolkit.
How much does web scraping cost per 1,000 pages?
With your own scraper, bandwidth for 1,000 HTML-only pages (about 100 KB each) costs $0.10 at $1/GB and $0.40 at the $4/GB market median; 1,000 full browser renders (about 2 MB) cost $2–$16. Managed APIs range from about $0.13–$0.28 per 1,000 plain pages to $36.75 for the hardest configuration on an entry plan. Use the cost calculator for your own numbers.
Is web scraping legal?
Scraping publicly available data is generally lawful in many places, but terms of service, copyright, personal-data rules (such as GDPR) and computer-misuse laws all apply. Some vendors refuse certain targets: Oxylabs publishes a list including banks, government and ticketing sites. This site is not legal advice; check with a lawyer for your use case.