The best SERP API providers in 2026

In mid-September 2025 Google stopped honouring the num=100 parameter. Scripts that used to pull the top 100 results in one request suddenly needed ten, and every provider that scrapes Google had to decide who eats that cost. Depending on the vendor, some of it landed on the customer’s bill. That’s a good reason to re-check what you’re paying for.

This list is for people who need search results as data: SEOs building their own rank tracking, developers putting SERP features into a dashboard, and anyone feeding live search results to an LLM tool. I’m writing from Singapore, where the location setting matters a lot to me. I want gl=sg results with a Singapore location, not US results wearing a Singapore label.

One limit up front. This is a desk review of six providers. I read the docs and pricing pages and compared how each one bills and shapes its output. I did not run a controlled accuracy benchmark across all six, so a table of prices can’t tell you which one returns the truest rankings. Prices are list prices and they move, so confirm on the vendor page before you put a card down.

How I picked

cost per 1,000 searches at the volume you will actually run. Plan headlines hide it, and a $75 plan and a $50 pack are different products once you divide.
depth. Since the num=100 change, the top 100 can mean ten billable requests, so how a vendor bills depth matters.
billing on failure. Whether a blocked or empty request still costs money.
location targeting. Country level at minimum, Singapore specifically. City level is a bonus, and you should test it against the vendor’s location list.
output. Structured JSON with organic positions and SERP features like featured snippets, People Also Ask and local packs, so nobody parses HTML.
legal posture. Who carries the risk when Google objects to scraping, which stopped being theoretical in December 2025.

The picks

In rough order of how often I’d suggest them, not by price.

DataForSEO

DataForSEO is where I’d start for anything that looks like rank tracking. It’s built for SEO work, so each result comes back with rank_group and rank_absolute. The first is the position among results of the same type, the second counts every block on the page, and the gap between them is the whole argument in what a ranking position actually means. Plenty of APIs give you one position number and leave you to work out what it counts.

You choose a delivery mode per request. The standard queue is cheapest but asynchronous: you post a task and fetch the result later. Priority is faster. Live answers in the same call and costs more than three times the standard rate. The same account also sells keyword data, backlinks and on-page audits, which helps if you’d rather not manage four vendors. The catch is glue code. The JSON is big and the task flow is more work than a single GET. Someone in the SerpApi camp will say the setup time isn’t worth it. At 5,000 first-page lookups a month it’s $75 against about $3 by list price, and I’d take the $3.

Pros:

roughly $0.60 per 1,000 SERPs on the standard queue, which undercuts every other paid option here at ordinary volumes
keyword, backlink and on-page data on the same account
pay as you go, no monthly plan to outgrow

Cons:

the standard queue is asynchronous, so it won’t suit anything a user is waiting on
verbose responses and more setup than the plug-and-play options

Pricing: from $0.0006 per SERP on the standard queue, $0.0012 on priority and $0.002 live. Pay as you go, with a $50 minimum top-up listed. Check how many results one “SERP” covers before you budget.

Link: dataforseo.com/apis/serp-api

SerpApi

SerpApi is what I’d hand a developer who wants a working call in ten minutes. One API covers Google, Bing, Baidu, Yahoo, DuckDuckGo, YouTube, Google Maps, Google Shopping and a long tail of other engines and marketplaces. The location parameter takes Google’s canonical location names, and there’s a locations endpoint to look them up, so “Singapore” resolves without guesswork. The docs and the in-browser playground are the friendliest of the six.

Google filed a lawsuit against SerpApi in December 2025 over how it gets past Google’s anti-scraping protections, and SerpApi advertises a legal shield for its customers. I haven’t followed the case since, so read the current status before you build a product on it. Every vendor here that scrapes Google carries some version of that risk. SerpApi is just the one with a docket number attached.

Pros:

widest engine coverage of the six
failed searches aren’t billed and cached results are free
SERP features already broken out into named JSON fields

Cons:

the priciest of the six per 1,000 searches, about $15 on Developer and $10 on Production
the lawsuit means you want a second vendor lined up

Pricing: free plan with 250 searches a month, Developer at $75 a month for 5,000 searches, Production at $150 for 15,000.

Link: serpapi.com/pricing

Serper

Serper (serper.dev) is what I’d use for a quick build, especially anything that passes search results to an LLM. The endpoints are simple: web, images, news, maps, places and shopping, and the JSON drops into a tool call without much cleanup. You get 2,500 free queries on signup, then buy packs. $50 buys 50,000 credits, which is $1 per 1,000, and bigger packs cost less per query. If you’re building the agent side of this, the AI tooling gets more coverage at AI Tool Gazette.

It looks like a small team and it’s Google-only in practice. I haven’t seen a legal indemnity or an enterprise support agreement offered, so assume the scraping risk sits with you and keep a fallback vendor in mind.

Pros:

$1 per 1,000 at the entry pack, with no monthly subscription
2,500 free queries to test with
small surface area, so integration is short

Cons:

Google only in practice, so no Bing or YouTube comparisons
thin docs and no enterprise support terms

Pricing: 2,500 free queries on signup, then $50 for 50,000 credits, cheaper per 1,000 on larger packs.

Link: serper.dev

Bright Data

Bright Data is a proxy company first, and its SERP API sits on that network. That’s the reason to look at it. If you need requests to come from a specific country or city and your volume is in the millions, that’s a problem it has spent years on. It lists Google, Bing, DuckDuckGo, Yandex and Baidu among others. You’re billed for successful requests only, and it handles CAPTCHAs and retries on its side.

I’ll take the unpopular side here: under about 100,000 searches a month I wouldn’t touch it. You’re paying for infrastructure you don’t need, and the plan structure takes longer to work out than the API does. Above that, ask for a quote, because the list price is a starting point.

Pros:

location targeting backed by a big proxy network
billed on successful requests only
retries and CAPTCHA handling stay on their side

Cons:

dearer than DataForSEO or Serper at small volume
plans and add-ons take effort to compare

Pricing: pay as you go is listed in the low single-digit dollars per 1,000 requests, cheaper on committed plans. Ask for a current quote.

Link: brightdata.com/products/serp-api

Oxylabs

Oxylabs is the other name that shows up once a procurement team gets involved. Its SERP scraping API also sits on a large proxy network and returns parsed Google and Bing results. The pitch is support: round-the-clock help and named account managers on the larger plans. I list it next to Bright Data because they solve the same problem. If nobody is asking you for an enterprise vendor on the invoice, skip both.

Pros:

parsed results, so no HTML handling on your side
human support and account managers on bigger plans
a free trial listed on the product page

Cons:

monthly plans, so small or spiky volume overpays

Pricing: plans from roughly $49 a month at listed prices, with the per-result rate falling on bigger tiers. Confirm on their site.

Link: oxylabs.io

Brave Search API

Brave Search API is the odd one out because it isn’t Google. It runs on Brave’s own independent index and it’s licensed access, so there’s no scraping question at all. The rankings won’t match Google’s, which makes it useless for Google rank tracking. Where it earns a slot is research, grounding for LLM answers, and a second opinion from an index that isn’t Google’s. It also supports goggles, Brave’s way of re-ranking results with rules you write yourself.

I haven’t compared its Singapore results against Google’s in any detail, so I won’t claim the two look alike. Run a few of your own queries side by side before you decide.

Pros:

licensed API, so no scraping risk
an independent index gives you a check on Google
simple per-request billing

Cons:

not Google, so no Google rank data and no Google SERP features

Pricing: billed per 1,000 requests. Check the current rate and free allowance on the page, because I’d rather not quote a number I can’t confirm today.

Link: brave.com/search/api

Comparison table

Entry prices are list prices as I have them. Treat them as a starting point and confirm before buying.

Provider	Entry pricing (list)	Primary strength	Primary weakness
DataForSEO	From $0.0006 per SERP, pay as you go	Lowest cost at ordinary volume, SEO data	Async standard queue, verbose JSON
SerpApi	Free 250 searches a month, then $75 for 5,000	Engine coverage and docs	Highest cost per 1,000
Serper	$50 for 50,000 credits	Simple JSON and low entry cost	Google only
Bright Data	Low single-digit dollars per 1,000	Location targeting at scale	Overkill for small volume
Oxylabs	Plans from roughly $49 a month	Support and account management	Monthly commitment
Brave Search API	Per 1,000 requests, see page	Licensed, no scraping risk	Not Google

How to choose

Start with volume, because it settles most of the decision. Under 250 searches a month, SerpApi’s free plan covers you, and Serper’s 2,500 signup queries are enough to test a project. From a few thousand up to a few hundred thousand, it’s DataForSEO or Serper. The split is whether you’d rather write glue code and pay less (DataForSEO) or pay a little more for a simpler call (Serper). Millions a month, or a buyer who wants an enterprise name, is Bright Data and Oxylabs territory.

Then price the depth you need, not the headline figure. If you only care about page one, the per-search numbers above are close to your real cost. If you track the top 100, multiply, and ask each vendor how a 100-result request is billed after the num=100 change. Ask what a blocked request costs while you’re there. Those two answers can move a bill more than the difference between $0.60 and $1 per 1,000.

Then the awkward part. Google’s terms of service limit automated access to its services, and the SerpApi suit shows Google will go to court over it. Vendors that scrape Google carry that risk, and some of it can land on you. A legal shield is a promise until someone tests it. This is not legal advice, and if the data goes into a paid product, spend an hour with a lawyer. Brave is one way around the question. Google’s own Custom Search JSON API is the other, and it has offered 100 free queries a day. It searches an engine you configure, not a mirror of what a searcher sees, so check what it offers before you plan around it.

Last, ask whether you need an API at all. If you’re checking a few hundred keywords a week, a rank tracker does it with less code, and I compared the options in the best SEO tools in 2026, tested. The APIs earn their place when you want raw SERP data to join to your own numbers, for example when you’re reading a competitor without copying them across a few hundred queries. If you do go the API route, test before you commit. Take 20 queries you know well, a few of them Singapore-specific, run them through two shortlisted vendors, and compare positions and SERP features to what a clean browser shows. Location handling is where I’d expect the differences to appear. More like this is on the blog index.

Verdict / top pick

My top pick is DataForSEO, for anyone whose job is SEO data at real volume. The price gap is too big to ignore, rank_group and rank_absolute answer the question that trips up most rank tracking, and pay as you go leaves nothing to outgrow. The cost is extra glue code, and I’d rather pay that once than a monthly premium forever.

If you want a working call in ten minutes, or Bing, YouTube and Maps behind one API, SerpApi is the pick, and you pay for the convenience. Serper suits agents and quick builds. Bright Data or Oxylabs only when volume or a procurement team demands it. Brave when the honest answer is that you don’t need Google at all.

Whichever you choose, keep the vendor behind a small wrapper in your code so switching takes an afternoon. This market moved twice in the past year, with the num=100 change and the Google lawsuit, and I’d expect it to move again.

Written by Xavier Fok

disclosure: this article may contain affiliate links. if you buy through them we may earn a commission at no extra cost to you. verdicts are independent of payouts. last reviewed by Xavier Fok on 2026-09-24.