Test
10,000 searches
$7.50
$0.75/1K searches
- Try every engine on your real queries
- Same API as the biggest pack
- Credits never expire
Quick answer
How can I use a SERP API for keyword research?
Search each seed keyword once and collect two lists from the JSON: people_also_ask questions and related_searches queries. Search those in turn for a second round. SerpCrypto returns both lists, plus the SERP features of each query, for one credit per search, from $0.30/1K searches as of October 2026. 200 seeds expanded twice is about 2,000 searches. No KYC; pay in crypto.
Keyword research with SerpCrypto
| Question ideas | people_also_ask: question, answer snippet, source link |
|---|---|
| Query ideas | related_searches: the suggestions at the bottom of the page |
| Intent signals | Which blocks appear: AI Overview, local pack, shopping, top stories, videos |
| Competition | The domains in organic_results for each query |
| Not provided | Search volume, keyword difficulty scores, autocomplete suggestions |
| Cost | 1 credit per search; 10,000 searches for $7.50 |
What can the results page tell me about a keyword?
It tells you what people want when they type it, as Google reads it. The blocks on the page are the clearest clue:
| Block on the page | What it suggests |
|---|---|
local_results (map pack) | Local intent: people want a place nearby |
shopping_results | Buying intent for a product |
top_stories | News intent: freshness matters |
inline_videos | A video answer is expected |
ai_overview | An informational query Google answers itself |
people_also_ask | Related questions worth answering on the same page |
The domains in organic_results show who you compete with, and whether the top results are big brands, forums or small sites like yours.
How do I expand a seed list without wasting searches?
Expand in rounds and stop early. Round one searches every seed; round two searches only the new questions and queries that look relevant; a third round is rarely worth it.
- Deduplicate new terms before searching them, ignoring case and punctuation.
- Filter by words that must appear, such as your product category, to stay on topic.
- Cap the number of searches per seed, so one broad seed cannot eat the budget.
The code below applies all three rules with a fixed search budget.
import os
import requests
API = "https://serpcrypto.com/api/v1/search"
HEADERS = {"Authorization": "Bearer " + os.environ["SERPCRYPTO_API_KEY"]}
def ideas(query, gl="us", hl="en"):
params = {"engine": "google", "q": query, "gl": gl, "hl": hl}
resp = requests.get(API, headers=HEADERS, params=params, timeout=60)
resp.raise_for_status()
data = resp.json()
found = [q["question"] for q in data.get("people_also_ask", [])]
found += [r["query"] for r in data.get("related_searches", [])]
return found
def expand(seeds, must_contain, budget=100, rounds=2):
seen = {s.lower() for s in seeds}
frontier, used = list(seeds), 0
for _ in range(rounds):
nxt = []
for term in frontier:
if used >= budget:
return sorted(seen)
used += 1
for idea in ideas(term):
key = idea.lower().strip(" ?")
if key not in seen and any(w in key for w in must_contain):
seen.add(key)
nxt.append(idea)
frontier = nxt
return sorted(seen)
if __name__ == "__main__":
for kw in expand(["standing desk"], must_contain=["desk"], budget=40):
print(kw)
Does SerpCrypto give search volume or keyword difficulty?
No. SerpCrypto returns results pages, not search volumes or difficulty scores. Volumes come from tools that measure demand, such as Google's own Keyword Planner inside Google Ads, or third-party SEO suites.
The two kinds of data work well together: a volume tool tells you how many people search a term, and the results page tells you what they expect to find and who already ranks. Autocomplete suggestions are not part of our engines either.
How do I turn questions into content?
Group the questions you collect by the page that should answer them. People Also Ask questions that share a topic belong on one page, each as a heading with a short, direct answer under it.
Then search each draft page's main query again after publication. If your page starts to appear in people_also_ask sources, in the AI Overview references or in the top results, the grouping worked. Our rank tracking guide shows how to follow that over time.
How many searches does a keyword research project use?
Count one search per term you look up. A project with 200 seeds and a second round on the 1,800 most relevant new ideas uses about 2,000 searches. Checking each idea in 3 countries triples that.
Research is usually a one-off or a quarterly job, so prepaid credits that never expire fit it well. 10,000 searches ($7.50) cover several projects of that size, and the balance waits for the next one.
Which pack suits keyword research?
One project usually fits in 10,000 searches ($7.50); agencies running research for many clients often start at 100,000 searches ($56.25), as of October 2026.
Starter
25,000 searches
$18.75
$0.75/1K searches
- A side project or an AI agent
- About 830 searches a day for a month
- Credits never expire
Build
100,000 searches
$56.25
$0.56/1K searches
- Daily rank tracking for 3,300 keywords
- One page per keyword per day for a month
- Credits never expire
In every plan: Every engine at 1 credit per search · Failed requests not charged · Credits never expire · Country, language, city and device targeting
Keyword research API: common questions
How many People Also Ask questions does one search return?
The questions Google shows on the page when it loads, usually a handful. Google reveals more when a person clicks a question; to go further, search the new questions themselves in a second round.
Can I research keywords in other languages?
Yes. Set gl and hl for the market, for example gl=fr&hl=fr for France. Questions and related searches come back in that language.
Is this the same as Google autocomplete?
No. Autocomplete suggestions are not part of the SerpCrypto engines. Related searches and People Also Ask come from the results page itself, after the search runs.
Can I see which SERP features a keyword triggers in bulk?
Yes. Search each keyword once and record which keys exist in the JSON, such as ai_overview, local_results or shopping_results. One search per keyword is enough.
How do I pay?
Top up from $25 in BTC, USDT, USDC, ETH, SOL, LTC and buy a pack from your balance. No ID, no selfie and no card.