Crawlbase pricing, and why your site's tier matters most

Crawlbase bills a falling ladder, from $3.00 per 1,000 successful requests down to $0.02, with no monthly fee, and sells optional $99 to $599 monthly blocks on top. Two things then move your bill more than the plan does: JavaScript rendering doubles it, and your target site’s difficulty tier multiplies it by anything from 1 to 20. Pricing, verified 2026-11-12.

Below I priced the ladder by hand, worked out 25K, 300K and 3M requests a month, showed how to read your site’s tier from a response header, and flagged the places where Crawlbase’s own pages disagree.

Short on time?

Here's my take

Trying it on your targets

Free credits

1,000 credits with no card, up to 5,000 after setup. Use them to read your site's tier before you budget.

Under about 143K standard requests a month

Pay as you go

No monthly fee: 25K costs $31.40 and 100K costs $76.40. Below 143K it beats the $99 Starter block.

About 143K to 400K a month

Starter, $99

300K costs $149 with overage, against $177.50 on pay as you go, because overage continues at the bracket you've already reached.

About 400K to 800K a month

Scale 500K, $199

Also the cheapest way to buy 300K JavaScript requests: $249 with overage, against $355 on pay as you go.

800K and up, including 2M to 3M

Scale 1M, $349, plus overage

3M costs $593.40, against $699 on Growth plus overage and $771.90 on pay as you go.

Hard targets (Complex, x2, or higher)

Zyte API as your benchmark

At x2, 300K requests cost $355 on pay as you go; at x20, $3,550. Price the same targets on Zyte's published tiers before committing.

Prices verified 12 Nov 2026. Sponsors cannot buy a verdict or a position.

Band of successful requests	Rate per 1K	Band cost	Running volume	Running total
First 1K	$3.00	$3	1,000	$3
Next 10K	$2.00	$20	11,000	$23
Next 100K	$0.60	$60	111,000	$83
Next 1M	$0.50	$500	1,111,000	$583
Next 10M	$0.10	$1,000	11,111,000	$1,583
Next 100M	$0.05	$5,000	111,111,000	$6,583
Next 1B	$0.04	$40,000	1,111,111,000	$46,583
Beyond	$0.02	–	–	–

Source: Rates from crawlbase.com/pricing, read 12 Nov 2026. Band costs and running totals are my arithmetic, for standard requests on a standard-tier site. Scroll sideways for more columns.

This page sits in the scraping APIs hub. If credits are new to you, what scraping API credits buy explains why a credit isn’t always a page.

How does the Crawlbase ladder work? #

It works like a tax bracket. Each band of successful requests in a month is billed at its own rate, so your first 1,000 cost $3.00 and the requests past 1.111 million cost $0.10 per 1,000. Pricing, verified 2026-11-12.

There’s a “$0” monthly fee and “no card to start”. Usage is “Billed monthly in arrears, for what you used”, and you can set a hard spend cap: “Cap the account or a single workspace and usage pauses before the bill passes it.” Pricing.

To price any volume by hand, take the running total at the last band edge below your volume and add the rest at the next band’s rate. For 25,000 requests that’s $23 + 14,000 at $0.60 per 1K = $31.40.

Crawlbase’s own examples are 100K for $76.40, 1M for $527.50 and 10M for $1,471.90, and my arithmetic reproduces all three. I priced different volumes below so this page doesn’t repeat the cheapest web scraping API comparison.

The steep part is the start. The first 11,000 requests average about $2.09 per 1,000, while everything from 1.111M to 11.111M costs $0.10. Small, steady jobs pay the most per request.

Is a subscription cheaper than pay as you go? #

Above about 143,000 requests a month, yes. Below that, the ladder costs less than the cheapest block: $83 + 32,000 at $0.50 per 1K = $99, which is exactly Starter’s price. Here are the four blocks, verified 2026-11-12. Pricing.

Two blocks share the name Scale, so I refer to them as Scale 500K and Scale 1M. Yearly billing is labeled “1 month free” and “8% off”, and every yearly price is exactly 11 times the monthly one, which works out to 8.3% off.

Past the block, the page says: “Usage past the block continues as pay as you go on the same ladder, at the bracket your volume already reached.” Turn pay as you go off and “requests pause at the block instead”, which gives you a hard stop. Pricing.

That overage sentence decides which block wins. Overage doesn’t go back to the top of the ladder: a Starter customer at 200K already sits in the $0.50 band, so the next 100K cost $50.

My arithmetic on published rates, standard requests on a standard-tier site. Starter hands over to Scale 500K at 400K, and Scale 500K hands over to Scale 1M at 800K. Growth never wins, because Scale 1M plus overage drops to $0.10 per 1,000 above 1.111M and stays $105.60 cheaper from 2M up.

So Growth at $599 is not the obvious pick for 3M that its size suggests.

One thing the page doesn’t settle. The pricing FAQ says blocks reset “with each monthly cycle” and pay as you go has “nothing to expire”, but rollover of unused block credits isn’t mentioned. On credits, the calculator counts 100K JavaScript requests as “200,000 credits spent”, so I’d expect a rendered request to draw two block credits.

What does JavaScript do to the bill? #

It doubles it. The pricing page says “A real browser render doubles the request” and “JavaScript rendering doubles any tier”, and the Smart AI Proxy page agrees: “A normal request uses one credit, a JavaScript request uses two.” Pricing, Smart AI Proxy, verified 2026-11-12.

You get two tokens. The Normal token is “Faster, cheaper”, and the JavaScript token is “Required to use page_wait, ajax_wait, scroll, and css_click_selector”. If you add javascript=true to a Normal key, “the request is billed as a JavaScript request.” Crawling API docs.

A few other things also bill as JavaScript:

• Scrolling: “The first 8 seconds (page load + scrolling combined) count as 1 request”, then “every full 5 seconds beyond that adds 1 more billed request, and a partial block is not billed.” So a 20-second scroll is billed as 3 JavaScript requests. Parameters.
• PDFs: “billed the same as a JavaScript-rendered request, since they share the headless pipeline”. Crawling API docs.
• Screenshots: “A full-page capture through the Crawling API spends two credits.” Pricing.

The cheapest habit is to send only the pages that need a browser through the JavaScript token. At 300K requests, rendering everything costs $355 on pay as you go, against $177.50 for none.

Which difficulty tier is your site in? #

You find out from the response, not from a list. The docs say the X-Domain-Complexity header (and the domain_complexity field) is “one of standard, moderate, or complex”, and that the tier “maps directly onto the pricing tier billed for the request”. Response docs, Crawling API docs, verified 2026-11-12.

The pricing page lists eight tiers, though, and says “Heavily defended sites carry a higher tier, applied automatically.” Its calculator shows only the first three, and there’s no public site-by-tier list. Pricing.

Multipliers from the pricing page; the dollar column is my pay-as-you-go arithmetic. Crawlbase’s 2024 explainer also named a “Fixed” tier, “a special category for domains that require a dedicated crawling setup with fixed pricing”, with LinkedIn as “the headline example”. Today’s pricing page doesn’t show it. Pricing explained.

My advice: spend some free credits on your real targets and log the header before you pick a plan. If a target comes back complex, treat ×2 as your floor, because the header can’t tell you whether you’re in one of the five tiers above it.

Does the multiplier apply before or after the ladder? #

After. The page says “Site difficulty and JavaScript rendering multiply a group’s ladder total, and the estimate above applies them exactly”, and elsewhere that the multipliers “scale a group’s whole ladder total”. Pricing, verified 2026-11-12.

The calculator on the same page confirms it. At 100K standard requests it shows $76.40. Switch JavaScript on and it shows $152.80 and “200,000 credits spent”; add Complex and it shows $305.60 and “400,000 credits spent”.

If 200,000 credits ran down the ladder first, they’d cost $127.50, not $152.80. So price your requests on the ladder, then multiply the dollar total: 300K JavaScript requests on a standard site cost 2 × $177.50 = $355.

So what will your month cost? #

Here are five jobs, priced on published rates and verified 2026-11-12. These are calculations, not invoices, and they assume every request succeeds.

The 25K row is where pay as you go shines: no subscription comes close, and you pay nothing in a quiet month. That steep first band costs you, though, as the effective rate is $1.26 per 1,000.

The 300K JavaScript row is the one to read twice. Six hundred thousand credits fit Scale 500K plus overage for $249, $106 less than pay as you go, as long as a rendered request draws two block credits, as the calculator’s “credits spent” count suggests.

At 3M, Scale 1M plus overage is $105.60 cheaper than Growth and $178.50 cheaper than pay as you go.

What counts as a billable request? #

Successful ones, with a wider definition of “successful” than most people expect. The docs bill a request only when cb_status is 200 or 207 and the target’s original_status is 200, 201, 301, 302, 303, 307, 308, 404 or 410. “Statuses you list in custom_success_codes are charged too.” Response docs, verified 2026-11-12.

So these are billed:

• 404 and 410 pages. A crawl of a stale URL list pays for every dead link.
• Empty pages. A 207 means “The target answered 200 with an empty body. Counts as a success and is charged.”
• Any status you add to custom_success_codes.

And these are free:

• Blocks and CAPTCHAs. A 503 means “Crawlbase recognized a block or CAPTCHA page”, even when the target returned 200. Response docs.
• Timeouts and target 5xx errors. “Failed requests (timeouts, blocks, 5xx from the target) are free”. Crawling API docs.
• Retries. “All retries against this API are free - only successful responses (cb_status: 200/207) count against your quota.” Crawling API docs.

One exception: on the async Crawler, webhook delivery retries are billed. Its docs say “Failed deliveries are charged” and “Every retry counts as a successful crawl for billing purposes”, so a flaky webhook endpoint costs real money. Crawler docs.

What do the free credits cover? #

You start with 1,000 credits and no card, and can reach up to 5,000. The pricing FAQ says “1,000 free credits and no card on file”, rising to 5,000 after setup steps it doesn’t name, while the hero says “Up to 5,000 requests free”. Pricing, homepage, verified 2026-11-12.

“Requests” and “credits” aren’t the same once JavaScript doubles a request: 1,000 credits are about 500 rendered pages. The ScraperAPI comparison page spells out the grant as 1,000 on signup plus onboarding steps. Crawlbase vs ScraperAPI. For other vendors’ free offers, see the free web scraping API roundup.

The credits come from one balance shared by every product: the Crawling API, the Enterprise Crawler, the Smart AI Proxy and Cloud Storage. Storage costs “Half a credit” to keep “one stored page for a month”. Pricing, Smart AI Proxy.

The Smart AI Proxy has no plans of its own. Bandwidth is “billed on requests rather than gigabytes”, and it draws the same “Optional subscriptions from $99 / mo.” The homepage also lists a Web MCP Server, but I found no price for it. Homepage.

Where Crawlbase’s own pages disagree #

Where they conflict, I trust the pricing page for rates and the response docs for what’s billed. Both are the most specific pages on their topic. The “optional” and “commitment-free” wording on the comparison page and blog does match the blocks: you never need one to use any feature.

What changed in 2026? #

The plan list was rebuilt more than once this year. A third-party tracker, PulseSignal (run by IA MARKETING SAS and structured with AI from the pricing page), logs plan sets replaced on three dates. PulseSignal, read 2026-11-12.

Crawlbase doesn’t publish an effective date for today’s model. PulseSignal’s own summary table still marks the old proxy and storage plans “Unchanged”, which conflicts with its log, so I treat the dates as a rough guide.

How does it compare at 300K requests? #

On easy pages, Crawlbase is mid-pack. On JavaScript pages it ties ScrapingBee. Each cell is the cheapest published way to cover 300,000 successful requests a month at list price, verified 2026-11-12.

Inputs: Zyte pricing, ZenRows pricing, ScrapingBee pricing, and my Zyte API, ZenRows and ScrapingBee pricing breakdowns. These are list prices for each vendor’s easiest case. Tiers and credit weights differ, so they aren’t the same work.

Crawlbase’s edge isn’t 300K of easy pages. It’s no monthly fee for small or spiky jobs (Zyte’s pay as you go shares that), free retries, and a rate that falls to $0.10 per 1,000 past 1.111M. ZenRows and ScrapingBee are monthly plans, so a quiet month still costs the plan price.

What about enterprise and very high volume? #

There are two separate “Talk to us” points. The pricing page asks “Need more than 2M credits a month?” next to a “Talk to us about enterprise” link and, at the far end of the ladder, “Crawling over 1B a month? Talk to us”. It says “Enterprise adds custom volume, SLAs and invoicing.” Pricing, verified 2026-11-12.

Throughput numbers are thin. The docs say only that “Concurrency limits scale with your plan”. The one figure I found is on the Zyte comparison page, in its Enterprise Crawler row: “3,000 RPM on standard plans, raised on request”, a vendor claim I couldn’t check. Crawling API docs, Crawlbase vs Zyte.

The async Crawler has published defaults. Pushes are “capped at 30 URLs/sec per token by default”, with 100K URLs per queue, 1,000,000 across crawlers, and a default concurrency of 20. Crawler docs.

What do customers say about the bills? #

The sample is small. Crawlbase’s Trustpilot profile, still filed under proxycrawl.com, had a TrustScore of 3.3 from 14 reviews when I read it. Trustpilot.

Support speed gets praise: “They responded immediately and it was unblocked in less than 24 hours” (July 2026). The price complaints are pointed. One two-star review (May 2026) called the pricing “unacceptable to me”, and a one-star review (June 2026) reported “Significant extra charges since September with no notification”, with no company reply shown.

Fourteen reviews prove nothing on their own. But surprise charges are what you’d expect from tiers applied automatically and billed in arrears, so I’d set the spend cap from day one.

Which plan should you pick? #

Trying it out: the free credits. Spend them on your real targets, log the X-Domain-Complexity header, and render only where you must.

Under about 143K standard requests a month: pay as you go. 25K costs $31.40, and a quiet month costs nothing.

About 143K to 400K: Starter at $99, with pay as you go left on for overage. 300K comes to $149, because overage continues at the $0.50 bracket.

About 400K to 800K, or 300K rendered: Scale 500K at $199. It’s also the cheapest fit for 300K JavaScript requests, at $249.

800K and up, including 2M to 3M: Scale 1M at $349 plus overage, which comes to $593.40 at 3M. That beats Growth plus overage ($699) and pay as you go ($771.90).

Hard targets at Complex or above: use Zyte API as your benchmark. A ×2 site can take 300K requests to $355, and Crawlbase won’t tell you in advance whether you’ll land in a ×9 or ×20 tier. Zyte publishes rates for each of its five tiers, so put the same targets through both before committing. If price is all you need on easy pages, the cheapest web scraping API comparison covers more options.

FAQ #

Does Crawlbase charge for failed requests? #

Mostly no. Blocks, CAPTCHA pages, timeouts and 5xx errors from the target are free, and so are retries. But the docs bill 404 and 410 responses and empty 200 pages (reported as 207). On the Crawler product, failed webhook deliveries are charged too. Response docs, Crawler docs.

How many free requests does Crawlbase give? #

1,000 free credits with no card, rising to up to 5,000 after setup steps the pricing page doesn’t name. Pages that still quote 10,000 free requests are out of date. Because JavaScript rendering doubles a request, 1,000 credits cover about 500 rendered pages. Pricing.

Does JavaScript rendering cost double on Crawlbase? #

Yes, according to the pricing page and the Smart AI Proxy page. One docs sentence says pricing is the same per successful response either way, but the same docs page bills javascript=true as a JavaScript request. I’d budget on double. Pricing, Crawling API docs.

Do unused Crawlbase subscription credits roll over? #

Crawlbase doesn’t say. The pricing FAQ says subscription blocks reset with each monthly cycle and doesn’t mention rollover, so I’d plan as if unused block credits are lost at the reset. Pricing.

Is ProxyCrawl the same as Crawlbase? #

Yes, ProxyCrawl is the old name. The docs note that the cb_status field was formerly named pc_status, the old ProxyCrawlAPI .NET package is marked deprecated as legacy, and Trustpilot still files reviews under proxycrawl.com. I couldn’t find an official rename date. Response docs, NuGet.

How do I know my site’s difficulty tier on Crawlbase? #

Read the X-Domain-Complexity header (or the domain_complexity field) on a response. It reports standard, moderate or complex, the tier you’re billed at. Spend some free credits on your real targets before you budget, because there’s no public list of sites by tier. Response docs.

Sources

22 sources, accessed 12 Nov 2026.

Official pricing, docs, changelogs and status pages 14

1. Crawlbase pricing, ladder, subscriptions and FAQ crawlbase.com · accessed 12 Nov 2026
2. Crawlbase homepage crawlbase.com · accessed 12 Nov 2026
3. Crawling API product page crawlbase.com · accessed 12 Nov 2026
4. Smart AI Proxy product page crawlbase.com · accessed 12 Nov 2026
5. Crawling API docs crawlbase.com · accessed 12 Nov 2026
6. Crawling API response docs (billing statuses, X-Domain-Complexity) crawlbase.com · accessed 12 Nov 2026
7. Crawling API parameters (scroll billing) crawlbase.com · accessed 12 Nov 2026
8. Crawler docs (limits and webhook billing) crawlbase.com · accessed 12 Nov 2026
9. Crawlbase pricing explained (Crawlbase blog, Dec 23, 2024) crawlbase.com · accessed 12 Nov 2026
10. Crawlbase vs Zyte (vendor page) crawlbase.com · accessed 12 Nov 2026
11. Crawlbase vs ScraperAPI (vendor page) crawlbase.com · accessed 12 Nov 2026
12. ScrapingBee pricing scrapingbee.com · accessed 12 Nov 2026
13. ZenRows pricing zenrows.com · accessed 12 Nov 2026
14. Zyte API pricing zyte.com · accessed 12 Nov 2026

GitHub code, issues and releases 1

1. Crawlbase GitHub organization github.com · accessed 12 Nov 2026

Review platforms 1

1. Crawlbase (proxycrawl.com) on Trustpilot trustpilot.com · accessed 12 Nov 2026

Other sources 6

1. ProxyCrawlAPI package on NuGet (deprecated) nuget.org · accessed 12 Nov 2026
2. PulseSignal change log for Crawlbase (third party) getpulsesignal.com · accessed 12 Nov 2026
3. apis.io Crawlbase plans (generated 2026-07-12) apis.io · accessed 12 Nov 2026
4. Prospeo on Crawlbase pricing (rival vendor) prospeo.io · accessed 12 Nov 2026
5. ScraperAPI's Crawlbase comparison (rival vendor) scraperapi.com · accessed 12 Nov 2026
6. Scrapfly's Crawlbase comparison (rival vendor) scrapfly.io · accessed 12 Nov 2026