ScrapingBee API review
Best for cheap rendered pages at volume
GET https://app.scrapingbee.com/api/v1/ Authorization: Bearer $SCRAPINGBEE_KEY # render_js defaults to true: 5 credits, not 1 # premium proxy 10 or 25, stealth 75, AI extraction +5 # a 429 is the concurrency limit, and it is not billed
The ScrapingBee API is one endpoint. You send it a URL and it fetches the page through rotating proxies, renders the JavaScript in a headless browser and hands back the HTML, a screenshot, or structured data from extraction rules you define. Search, Amazon, Walmart and YouTube have dedicated endpoints beside it, but the core is a single call.
Pricing is credits, and the number to understand is the multiplier. A plain fetch is 1 credit, but JavaScript rendering is on by default and costs 5, premium proxies cost 10 or 25, and stealth is 75. So the 75,000 credits on the $19 Hobby plan are 15,000 default calls, and the same credits buy 1,000 stealth pages.
The ceiling that binds is concurrency rather than credits. Every plan carries a lot of credits for its price, from 25 simultaneous requests on Hobby to 900 on the largest Enterprise plan, and a request over the limit is refused with a 429 that costs nothing.
The rest of the category is in the Sourcing APIs listing.
What one call costs, by configuration
ScrapingBee publishes the credit cost of every configuration, and only successful calls are charged. The Spb-cost response header tells you what each call actually cost, which is the easiest way to audit a pipeline.
| Configuration | Credits | At Hobby's $0.000253 a credit | Notes |
|---|---|---|---|
| Rotating proxy, no JavaScript | 1 | $0.00025 | render_js=false |
| Rotating proxy, JavaScript | 5 | $0.0013 | The default |
| Premium proxy, no JavaScript | 10 | $0.0025 | Residential IPs |
| Premium proxy, JavaScript | 25 | $0.0063 | |
| Stealth proxy, JavaScript | 75 | $0.019 | No stealth tier without JavaScript yet |
| AI extraction on top | +5 | +$0.0013 | ai_query or ai_extract_rules |
| Google Search API | 10, or 15 | $0.0025 | 15 with light_request=false |
Two settings save the most. Turn rendering off with render_js=false on any page that does not need it, which cuts a call from 5 credits to 1. And for blocked sites use mode=auto with a max_cost: it climbs the ladder from 1 credit to 75, charges only the configuration that succeeds, costs nothing if every tier fails, and stops at the cap you set.
Its reference notes list where its docs disagree. The custom_google option is priced at 15 credits in the CLI docs and 20 in the LangChain docs, and the current HTML API contract neither defines nor prices it. ScrapingBee's advice is to verify support and billing in the dashboard before using it, and that is the right advice for any parameter whose price you cannot find.
ScrapingBee API pricing, plan by plan
ScrapingBee API pricing is a monthly credit subscription, with concurrency rising alongside the credits and all prices exclusive of VAT. There is no pay-as-you-go option, and 1,000 free credits let you test before paying.
Hobby
- 25 concurrent requests
- 15,000 default rendered calls
- $0.000253 a credit
Freelance
- 50 concurrent requests
- 50,000 default rendered calls
- $0.000196 a credit
Startup
- 100 concurrent requests
- 200,000 default rendered calls
- $0.000099 a credit
Business
- 200 concurrent requests
- 600,000 default rendered calls
- Business + is $599 for 8,000,000
The curve is steep at the bottom. Startup costs about five times Hobby and carries more than thirteen times the credits, so the per-credit rate falls from $0.000253 to $0.000099. Above Business the Enterprise plans run from $999 for 14,000,000 credits to $5,799 for 120,000,000, with concurrency up to 900.
Credits are rarely the constraint here: 75,000 a month on Hobby is a lot of plain fetches. Concurrency is what sets how long a job takes, and it only rises with the plan. If a 50,000-page job has a deadline, the column to read is concurrent requests.
Concurrency is the limit, and a refusal is free
ScrapingBee API rate limits are a concurrency cap rather than a request rate: no per-minute limit is published for scraping, and each plan allows a fixed number of requests in flight at once. Exceed it and the call returns 429, which ScrapingBee does not bill.
| Plan | Concurrent requests | Credits a month |
|---|---|---|
| Hobby | 25 | 75,000 |
| Freelance | 50 | 250,000 |
| Startup | 100 | 1,000,000 |
| Business | 200 | 3,000,000 |
| Business + | 400 | 8,000,000 |
| Enterprise 14 to 120 | 500 to 900 | 14,000,000 to 120,000,000 |
Status codes that cost nothing
Only successful calls are charged.
- 429: the concurrency limit. Retry when a slot frees up.
- 401: no credit left on the account.
- In auto mode, a total failure across every tier costs 0.
- The Spb-cost header shows what a successful call cost.
The other ceilings
Per call, not per account.
- A request times out at 140 seconds by default.
- Non-HTML downloads are limited to 2 MB a request.
- The usage endpoint allows 6 calls a minute and uses no concurrency.
- Stealth proxy drops several options, custom headers and cookies among them.
ScrapingBee API rate limits, then, come down to one number per plan, and the usage endpoint reports it live as max and current concurrency. A pipeline that reads that before it fans out will never meet the 429 at all.
How to create a ScrapingBee API key
A ScrapingBee API key comes with a free account and 1,000 credits, no card required. Send it as a Bearer token; the older api_key query parameter still works but is deprecated for new integrations.
The key, for code
One header, every endpoint.
- Authorization: Bearer <key> on every call.
- The HTML API accepts GET, POST and PUT; POST and PUT forward your body.
- The usage endpoint returns credits used and allowed, and concurrency.
- A CLI, LangChain tools, and Make, n8n and Zapier integrations use the same key.
The MCP server, for an assistant
Hosted, with the key in the URL.
- The endpoint is mcp.scrapingbee.com/mcp with api_key in the query string.
- No OAuth and no Bearer header, so the full URL is a secret.
- Tools cover page text and HTML, extraction, screenshots, search and ecommerce.
- Cursor and Claude Desktop connect through mcp-remote.
The one inconsistency is worth acting on. The API deprecates the key in a query string because URLs leak into logs, and the MCP server requires exactly that, so keep the MCP URL out of source control and out of any shared configuration, as ScrapingBee itself advises.
A first ScrapingBee API call
This is the call ScrapingBee's HTML API reference shows: the target goes in the url parameter and the key in the header. Add render_js=false when the page does not need a browser, and the call drops from 5 credits to 1.
GET https://app.scrapingbee.com/api/v1/ # ?url=<target> and optionally &render_js=false # what is left on the account: GET https://app.scrapingbee.com/api/v1/usage
Where the ScrapingBee API documentation lives
ScrapingBee's documentation is at scrapingbee.com/documentation, and it publishes a plain-text version of every API reference under /llms for agents, indexed by scrapingbee.com/llms.txt.
| Surface | Where | What it answers |
|---|---|---|
| Documentation | scrapingbee.com/documentation | Every parameter, with examples. |
| Index for agents | scrapingbee.com/llms.txt | Which API to use for which job, with links. |
| HTML API contract | scrapingbee.com/llms/html-api.txt | The credit ladder, status codes and auto mode. |
| Remote MCP | scrapingbee.com/llms/remote-mcp.txt | Endpoint, client setup and the tool list. |
| Reference notes | scrapingbee.com/llms/reference-notes.txt | Where its own docs conflict. |
The reference notes file is unusual and useful. Few vendors publish a list of where their own documentation disagrees with itself; ScrapingBee does, and it is the first thing to read before relying on any parameter whose price is not in the HTML API contract.
The MCP page lists ScrapingBee's MCP access alongside the others in this directory.
ScrapingBee on video
The section is here either way, because the absence is worth knowing: a vendor with no walkthrough of its own API is a vendor whose docs are the only way in.
We have not found a walkthrough of the API on ScrapingBee's own channel yet. ScrapingBee's own homepage, about page, blog and pricing page link to X and LinkedIn and to no YouTube channel, so there is no channel this directory can prove is ScrapingBee's. When ScrapingBee publishes one, or sends us one, it goes here.
A vendor video is the vendor's claim about itself, which is why it sits in its own section.
Strengths and limits
ScrapingBee is a cheap, predictable fetch API whose price depends almost entirely on one parameter you control.
What it does well
- A published credit cost for every configuration, and the cost of each call in a header.
- Only successful calls are charged; a 429 and a 401 cost nothing.
- Auto mode climbs the proxy ladder and stops at a max_cost you set.
- Lots of credits for the price: 75,000 for $19 and 1,000,000 for $99.
- Plain-text API contracts for agents, including a list of its own doc conflicts.
- A hosted MCP server, a CLI, and Make, n8n and Zapier integrations.
Where it falls short
- JavaScript rendering is on by default, so the default call costs 5 credits, not 1.
- Stealth pages cost 75 credits, which turns 75,000 credits into 1,000 pages.
- No pay-as-you-go plan, only monthly subscriptions.
- The MCP server puts the API key in the URL.
- It returns pages, not records about people or companies.
The pattern is a low price that stays low only while you choose the configuration deliberately. Set render_js and use auto mode with a cap, and ScrapingBee is among the cheapest pages in this directory; leave the defaults on a defended site and the multiplier does the opposite.
Who should actually buy this
Buy ScrapingBee when a pipeline fetches many pages from many sites and the price per page matters more than anything else. Company websites, pricing pages and careers pages mostly render without stealth, and at Startup's rate a default rendered page costs about $0.0005.
Start on the 1,000 free credits and read the Spb-cost header on your real targets. That tells you which tier your sites need, and the tier decides the bill far more than the plan does.
Look elsewhere if you want clean Markdown for a model with no parameters to think about, structured records from a specific platform, or pay as you go. ScrapingBee is a fetcher; turning the page into a fact is the next step, and it is yours.
Before you build against it
Five things to settle first, each one read off ScrapingBee's own pages. The ticks are remembered in this browser, so you can come back to a half-finished integration.
Where the data comes from, and what you may do with it
The question that arrives late and expensively is who is allowed to hold the data. These are the fields it returns, the coverage it claims and the compliance it publishes.
| Provider | Fields returned | Coverage | Compliance | Redistribution |
|---|---|---|---|---|
| ScrapingBee | Firmographics, Intent and news, Web pages, Search results | Global | GDPR DPA | Not stated |
ScrapingBee states no redistribution position on its pricing or docs pages. Internally that never matters. Put the rows in something your customers can see and it becomes the question your legal team asks first.
Two practical notes. A compliance badge describes the vendor rather than your use of it, so a GDPR DPA does not give you a lawful basis for the outreach you send. And the vendor runs its own opt-out, which means a person who asks you to delete their record has to be removed from your copy as well as from theirs.
ScrapingBee against the obvious rival
Same rate cards, read side by side, with the arithmetic done on a worked volume rather than left to you: Apify vs ScrapingBee and Firecrawl vs ScrapingBee. Every pairing in the directory is on the compare hub.
Apify vs ScrapingBee
CompareOne endpoint priced in credits that scale with how hard the page fights, against a platform priced by the run.
Firecrawl vs ScrapingBee
CompareBoth render the page for you. One returns Markdown at a credit a page; the other returns HTML at five credits a page, and its credits cost far less.
Calling these APIs one by one? Run them from one Clay table.
Clay runs a waterfall across 150+ data providers from one table, without a separate contract with each. A lookup that finds nothing costs no data credit, and the free plan includes 500 actions a month.
Frontal is a Clay Elite Studio Partner.
Nothing worth quoting yet
We searched Reddit for threads about the ScrapingBee API and have found none worth quoting yet. When one turns up, it goes here, and nothing on this page leans on an unquoted paraphrase.
ScrapingBee API, answered
Is the ScrapingBee API free?
1,000 free API credits to try it, no card. Paid use starts at $19/mo.
How much does the ScrapingBee API cost?
ScrapingBee starts at $19/mo. Its unit rate is $0.00025 per credit. It bills by subscription or by credit pack, with a free tier. Read from ScrapingBee's own pricing page on 24 October 2026.
What are the ScrapingBee API rate limits?
Concurrency per plan, 25 to 900 requests at once. No per-minute cap published for scraping.
How do I get a ScrapingBee API key?
A ScrapingBee API key comes with a free account and 1,000 credits, no card required. Requests authenticate with bearer API key; the api_key query parameter is deprecated except for MCP. The current steps are in ScrapingBee's own documentation, which is authoritative on anything that has changed since 24 October 2026.
Does ScrapingBee have an MCP server?
Yes. Hosted, with the key in the URL. It is listed with the others on the MCP page.