100,000 ScraperAPI Credits Get You Only 4,000 Google Searches — What to Use Instead
ScraperAPI's pricing page leads with big, friendly numbers. The $49 plan gives you "100,000 API credits." It looks like 100,000 requests. For most scraping that is roughly true. For Google search, it is not even close.
A Google SERP costs 25 credits per request. So the $49 plan is really 4,000 Google searches, and the real price is about $12.25 per 1,000 — one of the most expensive ways to get Google data on the market. I read ScraperAPI's own credit documentation and worked the math out tier by tier. Here is the honest version, with the good parts kept in. If you just want the short answer on who is cheapest, the broader three-way SERP API cost comparison covers that; this post is the deep dive on the credit system itself.
The One Number That Changes Everything
ScraperAPI prices every request in "credits," and the credit cost depends on what you are scraping. A plain page is cheap. A search engine results page is not.
From ScraperAPI's official credit table, a request to Google or Bing falls into the "SERP" category, which is billed at 25 credits. Structured parsing (their autoparse option, which returns clean JSON instead of raw HTML) is included for free, so a parsed Google result is 25 credits flat — not 25 plus extras.
That single multiplier is the whole story. Divide any credit total by 25 and you get the number of Google searches you can actually run:
| Request type | Credits each | What 100,000 credits buys |
|---|---|---|
| Plain page (flat fetch) | 1 | 100,000 requests |
| E-commerce product page (Amazon) | 5 | 20,000 requests |
| Google / Bing SERP | 25 | 4,000 searches |
| Social profile (LinkedIn) | 30 | 3,333 requests |
So the same 100,000-credit bucket is 100,000 plain pages or 4,000 Google searches. The pricing page is technically correct — you do get 100,000 credits — but if your workload is search, the number that matters is 25× smaller.
ScraperAPI's Credit Table, Decoded
Credits stack with optional parameters, which is where surprise bills come from on non-search scraping. For Google you usually do not need these because the parsed SERP endpoint already handles the hard part, but it is worth seeing the full menu:
| Parameter | Extra credits | When you need it |
|---|---|---|
render=true (JavaScript) | +10 | JS-heavy pages |
premium=true (premium proxies) | +10 | Harder targets |
ultra_premium=true | +30 | The hardest anti-bot sites |
premium + render | 25 total | Protected, JS-heavy pages |
ultra_premium + render | 75 total | The worst case |
autoparse, country_code, device_type | 0 | Free add-ons |
This is the part reviewers complain about most: a job you budgeted at "1 credit each" can quietly become 25 or 75 once you turn on JavaScript rendering and premium proxies to get past a block. For Google search the 25-credit SERP rate already bundles what you need, so the danger zone is really the general scraping side — but it is the reason "how many requests do I actually get" is never the number on the box.
What 100,000 Credits Actually Buys
Think of credits as a separate currency, priced at 25 to the Google search. Here is what each paid plan converts to once you do that exchange:
| Plan | Monthly | API credits | Google searches (÷ 25) |
|---|---|---|---|
| Free | $0 | 1,000 / mo | 40 searches |
| Hobby | $49 | 100,000 | 4,000 |
| Startup | $149 | 1,000,000 | 40,000 |
| Business | $299 | 3,000,000 | 120,000 |
| Scaling | $475 | 5,000,000 | 200,000 |
The free tier is the clearest tell. "1,000 free credits a month" sounds generous until you realise it is 40 Google searches — barely enough to test that your code parses the JSON. The seven-day trial of 5,000 credits is 200 Google searches before it expires.
The Real Cost Per 1,000 Google Searches
Now convert plan price into a per-search cost. The formula is simple: (plan price ÷ credits) × 25 × 1,000.
| Plan | Cost per credit | Cost per Google search | Cost per 1,000 |
|---|---|---|---|
| Hobby ($49) | $0.00049 | $0.01225 | $12.25 |
| Startup ($149) | $0.000149 | $0.003725 | $3.73 |
| Business ($299) | $0.0000997 | $0.00249 | $2.49 |
The entry-plan figure of $12.25 per 1,000 is not just my arithmetic. An independent Google SERP benchmark by Scrape.do (a competitor, so read it with that in mind) measured the same $12.25 and noted "at 25 credits per SERP request, that provides 4,000 queries." The per-credit price improves a lot at scale — $2.49 per 1,000 on the Business plan is reasonable — but you have to commit $299 a month and a million-plus credits to get there. At hobby volume, you are paying premium-tier money for Google data.
Do the Credits Roll Over?
No. Credits reset each month and do not carry forward. ScraperAPI's plans and billing documentation describes a fixed monthly credit allotment with no rollover provision, and reviewers on G2 list "wish credits rolled over" among the common gripes.
Practically, that means a use-it-or-lose-it problem. If your traffic is spiky — a rank-tracking run on the 1st, quiet for three weeks — you have to size the plan for the peak and waste the rest. Pay-as-you-go providers avoid this entirely, which is part of why credit-bundle pricing draws so much friction. We unpack that trade-off in the SERP API pricing comparison.
The Part ScraperAPI Gets Right
An honest teardown keeps the good parts in, and there is a real one here: ScraperAPI does not charge you for failed requests. Their status-code documentation states you are only billed for successful requests (200 and 404). It retries a failing request for up to 70 seconds, and if it still cannot get through it returns a 500 and charges nothing.
That is genuinely customer-friendly, and it is more than some proxy-based providers offer. The only catch is latency: a request that quietly retries for tens of seconds before giving up can stall a pipeline if your client timeout is set too low. ScraperAPI's own advice is to allow a 70-second timeout, which tells you how long a hard case can take.
The Catch That Isn't On the Pricing Page
Price is only half of value. The other half is speed and how much data you actually get back, and this is where the 25-credit Google endpoint looks weakest.
In Scrape.do's eight-provider Google benchmark, ScraperAPI was the slowest of the lot at about 11.2 seconds per SERP — roughly six times slower than the fastest providers in the same test — while also being the most expensive. It returned a 100% success rate on Google, so reliability on search is not the issue. The data depth is. ScraperAPI captured 0% of AI Overviews in that test, plus no sitelinks, no publication dates, no video carousels, and no news results. If your use case depends on AI Overview text — increasingly common for SEO and AI-visibility work — ScraperAPI's parsed Google output simply will not contain it. (If that is your goal, see how to extract Google AI Overviews via API.)
One number to read carefully: a separate Scrapeway benchmark from June 2026 put ScraperAPI's success rate at just 34% — but that test covered hard, anti-bot-protected general sites (Zillow, LinkedIn, Instagram and the like), not Google. On Google search specifically, reliability is fine. The honest framing is: solid on Google reliability, but slow, expensive, and thin on rich SERP features.
How ScraperAPI's Google Cost Compares
Put the entry-tier cost next to the SERP-specialized providers and the gap is hard to ignore. All figures are the effective cost per 1,000 Google searches at an entry plan:
| Provider | Entry pricing | Per 1,000 Google searches |
|---|---|---|
| ScraperAPI | Hobby $49 (25 credits/SERP) | $12.25 |
| SerpApi | Developer $75/mo | $15.00 |
| Serper.dev | $50 starter pack | ~$1.00 |
| DataForSEO | Standard queue, $50 min deposit | ~$0.06 ($0.60 / 10K) |
| Serpent API | Pay-as-you-go, 10 free | from $0.003 ($0.03 / 10K) |
ScraperAPI and SerpApi sit at the top of the range; the SERP-specialized tools are an order of magnitude cheaper because they do one thing instead of scraping the whole web. For a head-to-head on the two cheap Google specialists, see the SerpApi vs Serper.dev benchmark.
When ScraperAPI Still Makes Sense
The credit model is expensive for Google, but ScraperAPI is not really a SERP tool — it is a general-purpose scraping platform, and for that job the pricing is reasonable. It is a fair pick when:
- Search is a small part of a bigger scrape. If you are already scraping product pages, listings, and protected sites, the 1-credit flat requests dominate your bill and the occasional 25-credit SERP is a rounding error.
- You need to get past Cloudflare, DataDome, or PerimeterX on non-search targets. That is ScraperAPI's core strength; if that is your problem, our guide on bypassing Cloudflare and DataDome explains where managed bypass earns its price.
- You value not being billed for failures and can tolerate higher latency on hard pages.
What it is not good at is being a cheap, fast, feature-rich Google SERP API. If that is all you need, you are overpaying by roughly 10× and getting no AI Overview data for it.
How to Pay Less for Google SERP Data
Three practical moves if Google search is your actual workload:
- Use a SERP-specialized API, not a general scraper. Tools built only for search skip the credit-multiplier model and price per search, which is where the 10× savings come from.
- Prefer pay-as-you-go over credit bundles if your volume is spiky. No rollover means you pay for the peak every month; usage-based billing only charges for what you run. The true cost of a Google scraper breaks down where the hidden money goes.
- Check AI Overview coverage before you commit. If you need it and your provider returns 0%, no price is low enough — you are buying incomplete data.
Get Google SERPs Without the Credit Math
Serpent API prices per search, not per credit — Google results with full AI Overview text and source citations, 10 free searches to test, and no monthly minimum or expiring credits. Pay-as-you-go from $0.03 per 10,000 searches on the Scale tier, a fraction of ScraperAPI's 25-credit rate.
Get Your Free API KeyExplore: SERP API · Pricing · Playground
FAQ
How many credits does a Google search cost on ScraperAPI?
A Google or Bing SERP request costs 25 API credits — the base rate for the SERP category in ScraperAPI's credit table, with structured parsing included free. So on the $49 Hobby plan, 100,000 credits equal 4,000 Google searches, not 100,000.
What does ScraperAPI really cost per 1,000 Google searches?
About $12.25 on the Hobby plan, $3.73 on Startup ($149), and $2.49 on Business ($299). Divide any credit total by 25 to get the real Google-search count, then divide the plan price by that.
Do ScraperAPI credits roll over to the next month?
No. Credits reset monthly and do not carry forward, so unused credits are lost. That makes overprovisioning costly for spiky or seasonal workloads.
Does ScraperAPI charge for failed or blocked requests?
No. Per their docs, you are only billed for successful requests (200 and 404). A failed request is retried for up to 70 seconds and, if it still fails, returns a 500 with no charge. The trade-off is latency on hard pages.
Does ScraperAPI return Google AI Overviews?
No. In Scrape.do's independent eight-way benchmark, ScraperAPI captured 0% of AI Overviews and no sitelinks, publication dates, video carousels, or news results. Its parsed Google output covers organic results, People Also Ask, and related searches only.
What is the cheapest ScraperAPI alternative for Google SERPs?
For Google data specifically, SERP-specialized APIs win on price: Serper.dev around $1 per 1,000, DataForSEO around $0.60 per 10,000, and Serpent API from $0.03 per 10,000 with 10 free searches and no monthly minimum. ScraperAPI is better kept for general-purpose scraping of anti-bot-protected, non-search sites.
