Give your AI agents fresh Google search results
Let an LLM agent or RAG pipeline search Google and get compact JSON back: organic results with real links and snippets, People Also Ask, and the AI Overview when Google shows one.
What gets in the way
A model only knows what it was trained on. Questions about prices, news or anything recent need a live search, and scraping Google yourself means handling blocks, layouts and parsing.
Agents work best with clean, structured input: titles, snippets and real URLs they can cite or fetch next, not raw HTML.
How CrawlerData solves it
Expose a search tool to your agent that calls the Google Search endpoint on the live queue and returns organic_results.
Google Search APILinks are resolved to the destination site, so the agent can cite them or fetch the page next.
Google Search APIAdd Google News for recent events, and Google Shopping for current product prices.
Google News APIai_overview, people_also_ask and related_searches give extra context and follow-up queries when Google shows them.
Google Search API
One request, structured data back
Pass organic_results (title, link, snippet) into the model's context. ai_overview is included when Google shows one.
curl -X POST https://api.crawlerdata.com/v1/jobs \
-H "Authorization: Bearer YOUR_API_KEY" \
-H "Content-Type: application/json" \
-d '{"endpoint":"google-search","params":{"query":"best crm software","country":"us","device":"desktop"},"queue":"live"}'const res = await fetch('https://api.crawlerdata.com/v1/jobs', {
method: 'POST',
headers: { Authorization: 'Bearer YOUR_API_KEY', 'Content-Type': 'application/json' },
body: JSON.stringify({
"endpoint": "google-search",
"params": {
"query": "best crm software",
"country": "us",
"device": "desktop"
},
"queue": "live"
}),
});
const { job, result } = await res.json();import requests
res = requests.post(
"https://api.crawlerdata.com/v1/jobs",
headers={"Authorization": "Bearer YOUR_API_KEY"},
json={
"endpoint": "google-search",
"params": {
"query": "best crm software",
"country": "us",
"device": "desktop",
},
"queue": "live",
},
)
data = res.json(){
"organic_results": [
{
"position": 1,
"title": "If You Had to Pick the Best CRM for Small Business Today, ...",
"link": "https://www.reddit.com/r/CRMSoftware/comments/1vmy4xb/if_you_had_to_pick_the_best_crm_for_small/",
"snippet": "I've spent the last few days reading reviews and watching videos, and somehow I'm even more undecide…"
},
{
"position": 2,
"title": "Enterprise CRM Software: Salesforce vs HubSpot ...",
"link": "https://www.alphabold.com/crm-software-solutions-for-large-companies/",
"snippet": "Sep 15, 2026 — This blog explores the top three CRM enterprise software solutions for large business…"
}
],
"ai_overview": {
"text": "The best CRM software depends on your business size and goals, with HubSpot CRM ranked as the top ch…",
"sources": [
{
"title": "HubSpot CRM",
"link": "https://www.hubspot.com"
},
{
"title": "Solutions Review",
"link": "https://solutionsreview.com/crm/2026/01/05/best-crm-companies-to-know-about/"
}
]
},
"people_also_ask": [
{
"question": "Will CRM be replaced by AI?"
},
{
"question": "What is the #1 CRM platform?"
}
]
}Long values are shortened in this example.
Excerpt of the result object, with fields and items trimmed. Full response on the Google Search API page.
A worked example
Every endpoint costs 1 credit per request. Here is what a typical monthly workload costs with whole credit packs.
5,000 agent searches a day × 30 days
= 150,000 requests a month
1 credit per request, so 150,000 credits a month.
Buy one month at a time
3 × Starter (150,000 credits) for $150.00
$150.00 / month
$0.0010 per search
Buy six months at once (best)
900,000 credits needed: 2 × Standard (1,000,000 credits) for $750.00
$125.00 / month
$0.0008 per search, 100,000 credits left unused after six months
Packs are bought whole, like in the pricing calculator. Credits are valid for 6 months, so buying ahead can move you to a cheaper pack. Failed requests are not charged. Prices exclude sales tax/VAT.
Limits to plan for
- A live request searches Google at request time and usually takes 10 to 120 seconds. Allow for that in your agent's tool timeout.
- Snippets are short. For full page content, your agent needs to fetch the linked page itself.
- Results are what Google shows for the country, language and device you set, at that moment.
- ai_overview is present only when Google shows an AI Overview for the query.
Questions
Can my agent call the API directly?
Yes. It is a single POST to https://api.crawlerdata.com/v1/jobs with a Bearer API key. Keep the key on your server, not in client-side code.
Is the response small enough for a context window?
The response is structured JSON. Pick the fields you need, such as title, link and snippet, before passing them to the model.
Which queue should an agent use?
The live queue, which returns the result in the same response. The fast and regular queues return a job ID to poll or a webhook, and cost the same.
Do failed searches use credits?
No. Failed requests are not charged.
Related solutions
SEO rank tracking
Organic positions and SERP features, desktop or mobile, per keyword and country.
See how it works
News and brand monitoring
Google News mentions of your brand, competitors and topics on a schedule.
See how it works
APIs used
Google Search API
Organic results with real destination links, ads, AI Overview, People Also Ask, local pack, top stories, knowledge panel and related searches. Desktop or mobile.
from $0.30 / 1,000 requests
Google News API
News articles with source, date, snippet and link.
from $0.30 / 1,000 requests
Google Shopping API
Search Google Shopping and get every product with price, merchant, rating and a product ID you can use to fetch all sellers.
from $0.30 / 1,000 requests
Try it on your own data
Get 50 free credits when you sign up with your work email and verify it. Failed requests are never charged.