Cost per 1,000 rendered pages
$1.27 on Hobby, $0.37 on Business+
With classic proxies and JavaScript on, the default request, bigger plans are more than three times cheaper per page. Freelance is $0.98 and Startup about $0.50.
ScrapingBee Pricing in 2026: ScrapingBee Plans, API Credit Costs and Real Cost per Page
9 min read
ScrapingBee costs $19 to $599 a month for 75,000 to 8,000,000 API credits. The catch is that a request costs anywhere from 1 credit to 75, depending on whether the site needs JavaScript, premium proxies or stealth mode. The same $49 Freelance plan buys 250,000 plain requests or about 3,333 stealth ones. Blocked requests are free; a 404 is not.
figures taken from scrapingbee.com/pricing and ScrapingBee's credit cost documentation
Robot console
Standing by Running · s Complete · rows
1 · Pick a target
3 · Fields to extract
Agent log
Crawl graph
Extracted data · rows
Want this data fresh every morning, without lifting a finger?
The plans
ScrapingBee plans and what each one includes
Five self-serve plans, all billed monthly, all excluding VAT. Every plan gets JavaScript rendering, rotating
and premium proxies, geotargeting, screenshots, extraction rules, the Google search API and the dedicated
scraper APIs. What changes is the credit pool, the number of requests you can run at once and, from Business
up, support and team features.
| Plan | Price a month | API credits | Concurrent requests | Support and team |
|---|---|---|---|---|
| Free credits | $0, once | 1,000 | n/a | No credit card |
| Hobby | $19 | 75,000 | 25 | Standard |
| Freelance | $49 | 250,000 | 50 | Standard |
| Startup | $99 | 1,000,000 | 100 | Standard |
| Business | $249 | 3,000,000 | 200 | Priority email, account manager, team management |
| Business+ | $599 | 8,000,000 | 400 | Priority email, account manager, team management |
| Custom | Quote | More credits or concurrency | Negotiated | Talk to sales |
Hobby is the cheapest way in. Concurrency is worth a look before you pick by credits alone: at 25 parallel requests and a few seconds per rendered page, Hobby moves through a large queue far slower than Business at 200. You can cancel any plan from the dashboard at any time.
The meter
How ScrapingBee credits work
A credit is not a page. Each request is charged by the options it uses, and the default request already costs
5, because JavaScript rendering is on unless you turn it off.
| HTML API request | Credits |
|---|---|
| Classic proxy, no JavaScript | 1 |
| Classic proxy, JavaScript rendering (default) | 5 |
| Premium proxy, no JavaScript | 10 |
| Premium proxy, JavaScript rendering | 25 |
| Stealth proxy (JavaScript always on) | 75 |
| AI extraction query, added to any of the above | +5 |
| A Google URL, classic or premium | 15 |
| Dedicated API request | Credits |
|---|---|
| Amazon, light or normal | 5 or 15 |
| Google or Walmart, light or normal | 10 or 15 |
| YouTube | 5 |
| Fast search | 10 |
| ChatGPT or Gemini | 15 |
| Agentic search | 3,750 |
ScrapingBee bills a request as successful when it returns 200, 404 or 410. Blocks and timeouts cost nothing, which is fair. Note the 404, though: if your URL list includes products that were delisted, every dead link still takes its credits. And you do not choose the mode by preference. You start cheap, the site blocks you, and you step up to premium or stealth until the page comes back, so the target site effectively picks your price.
Real volume
How many pages each ScrapingBee plan really buys
Same plans, divided by the credit cost of each request type. This is the table to read before you pick a
plan, because the gap between columns is up to 75 times.
| Plan | Classic, no JS (1) | Classic with JS (5) | Premium, no JS (10) | Premium with JS (25) | Stealth (75) |
|---|---|---|---|---|---|
| Hobby $19 | 75,000 | 15,000 | 7,500 | 3,000 | 1,000 |
| Freelance $49 | 250,000 | 50,000 | 25,000 | 10,000 | 3,333 |
| Startup $99 | 1,000,000 | 200,000 | 100,000 | 40,000 | 13,333 |
| Business $249 | 3,000,000 | 600,000 | 300,000 | 120,000 | 40,000 |
| Business+ $599 | 8,000,000 | 1,600,000 | 800,000 | 320,000 | 106,666 |
Cost per 1,000 stealth pages
$19.00 on Hobby, $5.62 on Business+
Sites behind serious bot protection move you to stealth. At that point Freelance is $14.70 per 1,000 pages and Startup $7.43, fifteen times the rendered rate on the same plan.
Detail pages count twice
One row can be two requests
If the price is on the listing page but the SKU is on the product page, each row costs its share of the listing page plus a full detail request. Budget by requests, not by rows in your final sheet.
Real months
What ScrapingBee costs per month: three worked bills
Each bill uses the published plan prices and credit table. The first one is where ScrapingBee is clearly the
cheaper tool.
Worked bill 01
300 products daily, rendered
9,000 requests a month at 5 credits is 45,000 credits: Hobby at $19. That is well under WebRobot Launch at $79, provided you already have the parser, the scheduler and somewhere to store the history.
Worked bill 02
Same job, the site wants premium proxies
9,000 requests at 25 credits is 225,000 credits, so Freelance at $49. Add AI extraction to skip writing selectors and it becomes 30 credits a request, 270,000 in all, which no longer fits: Startup at $99.
Worked bill 03
2,000 listings a day behind bot protection
60,000 stealth requests a month at 75 credits is 4,500,000 credits. Business stops at 3,000,000, so this is Business+ at $599. On WebRobot, 60,000 records a month is Scale at $249, with hourly runs and change alerts included.
The pattern holds across vendors that meter requests. Easy sites are cheap and hard sites are expensive, and you only find out which one you have after you start. Our Zyte pricing breakdown shows the same effect with five difficulty tiers, and the Firecrawl pricing guide shows it with a 5-credit charge for structured output.
The other half
What ScrapingBee pricing does not cover
ScrapingBee sells the fetch: proxies, a headless browser and retries behind one endpoint. It does that well.
Everything around the fetch is a separate job.
Not included 01
Deciding which URLs to request
The API scrapes the URL you send it. Walking a category, following pagination and collecting detail links is your code, and every page it walks is another billed request.
Not included 02
Parsing and fixing parsers
By default you get HTML back. Extraction rules and AI queries help, but when a site redesigns, the rules stop matching and the fix is on your side.
Not included 03
Scheduling and history
There is no built-in daily run or record of yesterday's prices. A cron job, a database and a diff are how you turn fetches into a price history.
Not included 04
Delivery to the people who use it
No-code integrations with Make, n8n and Zapier exist, but a Google Sheet that updates every morning and a Slack alert when a competitor drops a price are workflows you assemble.
For a team with a developer, that is fine, and ScrapingBee's price is honest for what it does. For a team without one, the real budget line is the person who writes and maintains the pipeline. If that is you, price it properly: either hire a developer to build and look after the scraping pipeline, or buy a tool that returns rows instead of HTML.
Comparison
ScrapingBee against other ways to pay for scraped data
Every vendor meters something different. Compare them by the unit you actually need at the end, which for most
business teams is a row in a table.
| Tool | What you pay for | Entry price | What you get back |
|---|---|---|---|
| ScrapingBee | Credits: 1 to 75 per request by mode | Hobby $19/mo | HTML, or JSON with extraction rules or AI |
| Zyte | Successful requests, priced by site difficulty tier | From $0.13 per 1,000 | HTML or automatic extraction |
| Firecrawl | Credits: 1 per page, 5 for structured JSON | Hobby $19/mo | Markdown or JSON |
| Diffbot | Credits: 1 per page, 25 per Knowledge Graph entity | Startup $299/mo | JSON in Diffbot's schema |
| Apify | Plan plus compute units and proxy usage | Starter $19/mo | Dataset from a prebuilt or custom actor |
| WebRobot | Records delivered, counted once per row | Launch $79/mo, $63 yearly | Named columns on a schedule, to a sheet, Slack, webhook or API |
If you are weighing the page-extraction APIs, our Diffbot alternative comparison covers Diffbot's credit plans and the Crawl gate, and Apify pricing explains compute units. For what a scraping project costs end to end, including people, read how much web scraping costs.
Decision
ScrapingBee or a ScrapingBee alternative
Pick ScrapingBee if
You have a developer and easy targets
Your team writes Python or Node, your target sites work on classic or premium proxies, and you want one endpoint that handles rendering and retries. At $19 to $99 a month it is one of the cheaper fetch APIs for that job.
Pick WebRobot if
You want rows, not a pipeline to build
You describe the data in a sentence and a hosted robot finds the pages, fills your columns and delivers them every day. Launch is $79 for 10,000 records; Scale is $249 for 100,000 with hourly runs, logins and alerts. See WebRobot pricing or run the demo on your own URL.
Teams that land here usually run one of two jobs: competitor price monitoring across retailers, or feeding their own product with the web scraping API, where the robot returns finished records instead of pages.
FAQ
ScrapingBee pricing questions
ScrapingBee has five self-serve plans billed monthly: Hobby at $19 for 75,000 API credits, Freelance at $49
for 250,000, Startup at $99 for 1,000,000, Business at $249 for 3,000,000 and Business+ at $599 for
8,000,000. Prices exclude VAT. Larger volumes are custom, and new accounts get 1,000 free credits with no
credit card.
Between 1 and 75 credits. A classic proxy request without JavaScript costs 1, with JavaScript 5, which is
the default. Premium proxies cost 10 without JavaScript and 25 with it. Stealth proxies cost 75. An AI
extraction query adds 5 credits. Google search requests cost 15, and the Amazon, Walmart, YouTube and
ChatGPT APIs cost 5 to 15.
Divide the plan's credits by the cost of your request type. Freelance at 250,000 credits buys 250,000
plain requests, 50,000 JavaScript-rendered pages, 10,000 pages on premium proxies with JavaScript or about
3,333 stealth requests. The mode a site forces on you matters far more than the plan name.
ScrapingBee charges only for successful requests, which it defines as responses with a 200, 404 or 410
status code. A blocked request or a timeout is free. A page that no longer exists and returns 404 is
billed, so a stale URL list still costs credits.
There is no ongoing free plan. New accounts get 1,000 API credits with no credit card, which is 200
JavaScript-rendered pages or about 13 stealth requests. After that the cheapest plan is Hobby at $19 a
month for 75,000 credits.
You can upgrade to a bigger plan for the current month and downgrade later, or renew your current
subscription early, which starts a new billing period with a fresh credit allowance. Either way you pay
again; there is no per-credit overage rate on the self-serve plans.
ScrapingBee is worth it for developers who write their own parsers and want proxies, rendering and retries
handled behind one endpoint, especially on sites that work with classic or premium proxies. It is poor
value for a business team that wants rows in a spreadsheet, because the parser, scheduler and delivery
are code you still write and maintain. The guide to
web scraping without getting blocked
covers why sites push you into the expensive modes.
Pay for rows, not requests
Describe the data in plain English. The robot finds the pages, fills your columns and delivers 10,000 records a month for $79.