upAPI vs ScrapingBee
Named operations with a published weight, not a credit meter on every page.
TL;DR: Is ScrapingBee or upAPI the better fit?
Pick ScrapingBee when you fetch pages you choose yourself and need credit-priced control over JavaScript rendering, premium proxies and geotargeting. Pick upAPI when the job is a named, structured result such as page metadata, a Maps search, a screenshot or Markdown, on one key with a free plan and a published weight per call.
How ScrapingBee charges
Monthly credit plans from $49 (250,000 credits) up to $3,599, with a request costing 1 to 75 credits depending on proxy type and JavaScript rendering, and 1,000 free credits to try it.
How upAPI charges
Free at $0 with 1,000 weighted units, then $29 for 100,000. Each operation carries one published weight, so a screenshot is 20 and a metadata read is 1, whichever proxy or browser runs behind it.
USD 29 / mo · USD 0.29 per 1,000 units · full pricing
Last checked 2026-10-07. ScrapingBee figures here were read from their own pages on that date. We do not restate a competitor’s number we have not seen ourselves, and where they publish none we say so; if this date has aged, check theirs before ours.
ScrapingBee and upAPI side by side
| Feature | ScrapingBee | upAPI |
|---|---|---|
| Pricing | Monthly credit plans: Freelance $49 (250,000 credits), Startup $99 (1,000,000), Business $249 (3,000,000), Business+ $599 (8,000,000), then Enterprise tiers from $999 to $3,599. A request is 1 credit with the classic proxy and no JavaScript, 5 with JavaScript rendering (the default), 10 with a premium proxy, 25 premium with JavaScript and 75 stealth with JavaScript. Prices exclude VAT. | Free at $0, then Starter $29 for 100,000 weighted units a month, Pro $99 for 500,000 and Scale $499 for 5,000,000. Each operation declares its own unit cost, and the quota is pooled across the catalog. |
| Free tier | 1,000 free API credits with no credit card. Their pages do not say whether the 1,000 recur monthly. | The Free plan: 1,000 weighted units a month across every public operation, with no trial clock. |
| Endpoints | One HTML API that takes any URL, plus dedicated APIs for Google, Fast Search, Amazon, Walmart, YouTube, ChatGPT and Gemini. | Named operations, each with its own request and response schema: Maps, search suggestions, page metadata, sitemaps, screenshots, PDF and more. There is no generic fetch-any-URL-as-HTML endpoint. |
| Output format | Page HTML, Markdown with return_page_markdown, screenshots, and AI extraction (ai_query, ai_extract_rules) at 5 extra credits. | JSON on every operation, shaped by its published output schema. The page-to-Markdown operation returns Markdown. |
| JavaScript rendering | Yes. render_js is on by default and costs 5 credits instead of 1; screenshots require it. | Only inside operations built on a browser: screenshot and HTML-to-PDF. Other operations have no render flag, and the page-to-Markdown operation reports a client-rendered page as an error rather than returning it empty. |
| Proxies | Automatic rotation on every request, premium residential and stealth proxies at higher credit costs, and country-level geolocation. | Included in the unit price. The exit is chosen per operation, and many operations accept your own proxy URL. |
ScrapingBee is a focused tool, and the comparison should be read that way. If the problem is "this page blocks me", their per-request controls are more expressive than anything in our catalog, and their credit schedule makes the cost of each proxy choice explicit.
The comparison changes when the job is "get these named facts from a handful of sources". A credit balance against one HTML endpoint leaves you writing and maintaining an extractor for every source, and the extractor is where the ongoing cost lives. ScrapingBee does offer AI extraction rules at 5 extra credits, which narrows the gap; upAPI’s named operations return their shape without a rule to write.
Entry price is the other difference: $49 against $0 and $29. If you are still deciding whether the data is useful, that gap matters more than the per-request rate.
Where ScrapingBee wins
- Any URL, with per-request control. Their HTML API takes a page you choose and lets you set JavaScript rendering, premium or stealth proxies and country geolocation on each request. upAPI has no operation like that.
- Dedicated APIs for Google, Amazon, Walmart, YouTube, ChatGPT and Gemini, which upAPI does not publish. Our Amazon product operation is gated private.
- Higher concurrency at volume: their plan table lists 50 on Freelance rising to 800 on the largest Enterprise plan, for heavy sustained scraping.
Where upAPI wins
- The result is already structured. A page-metadata, sitemap or Maps operation returns parsed JSON, so there is no parser for you to write and keep working.
- A lower way in. The free plan costs nothing and the first paid plan is $29, against $49 for their smallest.
- The price does not move with the proxy. A ScrapingBee request costs between 1 and 75 credits depending on proxy and rendering; an upAPI call costs its published weight.
- Agent access without a key in a URL. ScrapingBee’s remote MCP takes the key as a query parameter; its docs say it has no header or OAuth flow. upAPI’s hosted server signs you in through the browser over OAuth.
When to pick which
- ScrapingBee. You fetch arbitrary URLs you pick yourself and need per-request control over JavaScript rendering, proxy tier and country.
- ScrapingBee. You need their Google, Amazon, Walmart or YouTube APIs, which upAPI does not publish.
- upAPI. You want structured results from named sources, such as a Maps search or a page’s metadata, without writing and maintaining the extractor.
- upAPI. You are still testing whether the data is useful, and want a free plan with a published per-call weight before committing to a monthly spend.
- Both. A few stubborn pages need unblocking and the rest are named lookups. Use ScrapingBee for the first and upAPI for the second.
What switching actually costs
Replace the ScrapingBee request, an API key and a URL parameter, with a POST to `api.upapi.io/{slug}` with an `X-Api-Key` header and a JSON body, and delete the extractor if a named operation covers the source. Run both on the same inputs on the free plan first.
Questions about ScrapingBee and upAPI
How much does ScrapingBee cost compared with upAPI?
ScrapingBee sells monthly credit plans from $49 for 250,000 credits, with a request costing 1 to 75 credits depending on proxy type and JavaScript rendering (their pricing page and docs, read 2026-10-07). upAPI is free at $0 with 1,000 weighted units a month, then $29, $99 and $499. A credit and a weighted unit are different meters, so compare the price of the job you run, not the plan sizes.
Does ScrapingBee have a free plan?
It offers 1,000 free API credits with no credit card. Its pages do not say whether that allowance repeats each month. upAPI’s free plan is 1,000 weighted units a month, with no trial clock.
Does ScrapingBee charge for failed requests?
Only for successful ones. Their FAQ says they charge for requests returning a 200, 404 or 410 status code. With mode=auto, the docs say you pay only for the tier that succeeds.
Can upAPI replace ScrapingBee for scraping any URL?
No. upAPI has no operation that fetches an arbitrary URL and returns its HTML with render and proxy flags. The closest are page-to-Markdown, which reads any URL over plain HTTP without a browser, plus screenshot and HTML-to-PDF, which use a headless browser, and operations for page metadata and sitemaps.
Does ScrapingBee support MCP?
Yes. It hosts a remote MCP server at mcp.scrapingbee.com/mcp, and its docs pass the API key as a query parameter on that URL. upAPI’s hosted MCP server signs you in through the browser over OAuth, and a local server takes an upapi_ key.
Sources read on 2026-10-07: https://www.scrapingbee.com/pricing/, https://www.scrapingbee.com/documentation/, https://www.scrapingbee.com/faq/, https://www.scrapingbee.com/documentation/remote-mcp/, https://www.scrapingbee.com/llms.txt.
Other comparisons: Apify, Bright Data, ScraperAPI, Serper, ApyHub, RapidAPI, Firecrawl, ScreenshotOne.