Two of the most popular web scraping APIs, compared on the axis that actually decides the bill: how each one counts credits. Written by the team at WebScraping.AI — a competitor to both, which is exactly why we keep this factual and sourced.
You want a clean, well-documented general-purpose API with predictable feature-based multipliers, a Screenshot API, and high concurrency at entry tiers. Costs scale with the features you turn on (JS, premium proxies), not with which sites you scrape.
You want pre-parsed structured JSON from Amazon, Google, or Walmart without writing extraction code, or you operate at very large scale and can use its async scraper — and you're comfortable with per-domain credit surcharges.
Pricing and credit costs last checked July 2026 — both vendors change them; verify before deciding.
| ScrapingBee | ScraperAPI | |
|---|---|---|
| Entry plan | $49/mo · 250,000 credits (Freelance) | $49/mo · 100,000 credits (Hobby) |
| 1M-credit tier | $99/mo (Startup) | $149/mo (Startup) |
| Credit model | By feature: base 1, JS +, premium/stealth proxies up to 75 | By feature and domain: base 1, JS ~10, Amazon 5, Google/Bing 25, LinkedIn 30 |
| JS rendering | Yes (5 credits) | Yes (~10 credits) |
| Structured endpoints (parsed JSON) | Google Search API | Amazon, Google SERP, Walmart, more |
| AI extraction | Yes (ai_query/ai_extract_rules) | Yes (structured endpoints, LangChain/LlamaIndex) |
| MCP server | Yes (hosted) | Yes |
| Screenshot API | Yes | — |
| Async / batch at scale | Batch tooling | Async scraper, high thread limits on upper tiers |
| Official SDKs | Python, Node | Python, Node, PHP, Ruby, Java |
| Bills failed requests | No — successes only | No — successes only |
| Ownership | Oxylabs (acquired 2025) | Independent (owns Rainforest API since 2024) |
Sources: ScrapingBee pricing · ScraperAPI pricing
Both APIs advertise credit-based pricing, but the credits mean different things.
ScrapingBee multiplies by feature. A plain request is 1 credit; turning on JavaScript rendering or premium/stealth proxies raises the multiplier (up to 75 credits for the hardest targets). Your cost depends on what the target site requires, but not on which site it is. Estimating a job means knowing how many pages need JS and premium proxies.
ScraperAPI multiplies by feature and domain. A generic page is 1 credit, but Amazon costs 5, Google and Bing 25, LinkedIn 30 — before rendering multipliers. In exchange, those domains come back as pre-parsed structured JSON if you use the dedicated endpoints. Two jobs of identical size can produce very different bills depending on the target list.
The practical test: list your target domains first. If your list is mostly Amazon/Google/Walmart and you want parsed JSON, ScraperAPI's surcharges buy you real parsing work. If your list is broad or changes weekly, feature-based pricing is easier to predict.
On reliability: independent benchmarks (e.g. Scrapeway's monthly runs) have generally shown ScraperAPI with higher success rates but slower responses, and ScrapingBee faster but less consistent on some protected targets. Success rates shift month to month as anti-bot systems evolve — run your own pilot on your target list before committing either way.
This page is published by WebScraping.AI, which competes with both products — that's why every number above links to a source. For completeness, here's where we sit on the same axes: flat feature-based credits with no per-domain surcharges (1 plain / 5 JS / 10–25 residential / 50 stealth, +5 for AI extraction), a $29/mo entry plan with 250,000 credits, AI question answering and field extraction on any site, 7 official SDKs, and an MCP server. Failed requests are free — same as both tools above.
Is ScrapingBee or ScraperAPI cheaper?
It depends on your targets. At the 1M-credit tier ScrapingBee is $99/mo vs ScraperAPI's $149/mo, and ScrapingBee's entry plan includes 2.5x more credits at the same $49. But ScraperAPI's structured endpoints do parsing work you'd otherwise build yourself — for an Amazon- or Google-heavy pipeline, that can be worth the per-domain surcharge.
Which is better for scraping Amazon or Google?
ScraperAPI, if you want pre-parsed JSON: its dedicated endpoints return structured Amazon and Google data at 5 and 25 credits per request respectively. ScrapingBee offers a Google Search API but no Amazon parser; you'd extract with its AI rules or your own code.
Do ScrapingBee and ScraperAPI charge for failed requests?
No — both bill only successful requests. That's standard among the major scraping APIs (WebScraping.AI included).
Which has better AI and agent support?
Both ship MCP servers and AI extraction. ScrapingBee has ai_query/ai_extract_rules for plain-English extraction on any page; ScraperAPI leans on structured endpoints plus LangChain and LlamaIndex integrations. Neither has a decisive edge — pick by pricing model and target list.
Is there an alternative to both?
Several — this market is competitive. WebScraping.AI (our product) offers the same core stack with flat per-request pricing from $29/mo and no per-domain surcharges; Zyte, ZenRows, and Scrapingdog are also worth evaluating. Run a small pilot on your actual target sites before committing.
Try WebScraping.AI with 2,000 free API credits. No credit card required.