Title
The title as Google shows it.
Popular platforms
By use case
Start building
Organic web search results with URL, title, description, displayed URL and, when the page prints one, its date.
100 free credits. No card needed.
The title as Google shows it.
Position as Google gives it.
Displayed as Google gives it.
Its link, such as outstand.so.
What kind of account it is, such as organic.
Snippet as Google gives it.
Each result keeps its position, title and snippet, so a list of phrases becomes a table of the pages that hold each one, ready to check every week.
Outstand - Unified Social...
Social Media Posting API | 16...
date_posted takes hour, day, week, month or year, so a weekly digest can read only the pages printed in the last seven days.
Newsroom monitoring2
organic
Social Media Posting API |...
Results come from Yahoo Japan's search, not from google.com, and extra.provider says so. Organic results only: no ads, answer boxes or related searches.
yahoo_japan
organic
1
Sign up and get 100 free credits. No card needed.
Pass query (social media api).
Send back cursor until has_more is false.
1curl -G "https://api.socialscrape.dev/v1/google/search" \2 --data-urlencode "query=social media api" \3 -H "x-api-key: $SOCIALSCRAPE_API_KEY"1import requests2 3r = requests.get(4 "https://api.socialscrape.dev/v1/google/search",5 params={"query": "social media api"},6 headers={"x-api-key": KEY},7)8results = r.json()["data"]["items"]1const url = new URL("https://api.socialscrape.dev/v1/google/search");2url.searchParams.set("query", "social media api");3 4const res = await fetch(url, {5 headers: { "x-api-key": KEY },6});7const { data } = await res.json();{ "ok": true, "credits_charged": 1, "data": { "items": [ { "title": "Outstand - Unified Social Media...", "url": "https://www.outstand.so/", "position": 1, "displayed_url": "https://www.outstand.so", "snippet": "One API call posts to X...", ... }, ... ], "cursor": "2", "has_more": true } }
Ask your agent for Google-style web search results in plain words, and the SocialScrape MCP server makes the call.
Set up the MCP serverGoogle-style web search results, given a search.
1 credit a call; a repeat within 1 days is free.
Your key stays in your agent's settings.
18,000 credits a month
50,000 credits a month
150,000 credits a month
$27/mo
$73/mo
$209/mo
Over 1M credits a month? Talk to us. Or see full pricing.
Measured against real Google first pages for 18 searches, 5.5 of 8.7 URLs on average were also on Google's page (63%, from 2 to 8), and the first result was the same on 12 of 22 tries. The order, some URLs and some snippets differ.
Pages 1 to 10, about 10 results each. The cursor is the next page number, and page works the same way.
No. There is no region or language option, and a search for places near you follows our servers' location, not yours.
The results come from Yahoo Japan's search (search.yahoo.co.jp), not from google.com: the order, some URLs and some snippets differ from Google's own page. Measured against stored real Google first pages for 18 searches, 5.5 of 8.7 URLs on average were also on Google's page (63%, from 2 to 8; the first result was the same on 12 of 22 tries). Organic results only: no ads, answer boxes or related searches. date_posted takes hour, day, week, month or year. The cursor is the next page, 1 to 10 of about 10 results each; page works as the cursor. There is no region or language option, and searches for places near you follow our servers' location, not yours. When the search provider refuses every try the call fails and can be retried. Each result has up to 9 fields, and the sample fills position, title, url, displayed_url, snippet and kind.
query (required): Search terms.
date_posted: Only results from this window. One of hour, day, week, month or year.
invalid_params and not_found. None is charged, and the errors list says which to retry.
client.google.search(...) in Python and client.google.search() in TypeScript.