FlowRunner
PricingContact
Theme
Start Free

SerpApi

SEO

Run live, structured searches across 15 engines through SerpApi, including Google Search, Maps, News, Scholar, Bing, YouTube, and Amazon. Agents get fully parsed JSON results, re-read archived searches without spending new credits, and monitor account usage.

4 actions API key available
SerpApi website ↗ Platform Documentation ↗ Capability data verified 2026-07-27
The nightly rank-tracking schedule fires with the keyword list for the brand's core pages
Get Account reads searches used this month, remaining credits, and the hourly rate limit
The agent compares the batch size against remaining quota and the retention of yesterday's archived runs
If the batch would push usage past the month's budget line, the SEO lead decides whether to run it, trim the keyword list, or wait for the cycle reset
Search runs each keyword against Google with the canonical location and device profile
Titles, links, and positions append to the rank-tracking sheet, keyed to each search's archive ID
Position changes beyond the alert threshold post to the channel with the SERP feature that moved

What This Integration Enables

SerpApi's proposition is that search results are data, not pages: one Search action returns fully parsed JSON from Google Search, Maps, News, Images, Scholar, and Jobs, plus Bing, DuckDuckGo, Yahoo, YouTube, Amazon, eBay, Walmart, Baidu, and Yandex. For a FlowRunner agent that changes what a search is. It stops being a browser task and becomes a queryable input any workflow can branch on: a position, a price, a local pack entry, a news mention. The constraint that shapes every serious SerpApi workflow is the credit pool. Each live Search spends quota; Get Account, Search Locations, and archive reads do not. Well-built agents exploit that asymmetry, and FlowRunner's [human-in-the-loop](/concepts/human-in-the-loop) gate sits exactly where the spending decisions get large. - Track rankings and SERP features across engines, locations, and devices on a schedule - Pull structured product listings from Amazon, eBay, and Walmart searches for pricing work - Re-read archived searches by ID so analysis and retries never re-spend credits - Budget the monthly quota with pre-flight Get Account checks and human sign-off on large batches

Without FlowRunner

Rankings checked by hand Someone searches the brand's keywords in an incognito window and eyeballs the positions
SERP data trapped in HTML Scraping results means parsing markup that changes shape whenever the engine redesigns
Credit burn discovered at zero The quota runs out mid-month and nobody knows which workflow spent it

With FlowRunner

Structured results on a schedule Agents get parsed JSON positions, features, and listings from 15 engines without touching HTML
Every search re-readable Archive IDs let workflows re-read yesterday's results without spending a new credit
Quota as a managed budget Get Account runs before every batch, and a person approves any run that threatens the ceiling

Use Case Scenarios

Rank tracking that pays for itself in credits saved

An SEO team tracks 400 keywords nightly. The agent calls Get Account first, projects the batch against remaining quota, then loops Search with each keyword, canonical location from Search Locations, and a mobile device profile. Positions and archive IDs land in [Google Sheets](/integrations/google-sheets). When an analyst wants to investigate a drop the next morning, the agent re-reads the exact SERP with Get Search by ID instead of running a fresh query, so the investigation examines the same results the alert was based on and costs nothing.

Marketplace price watch across three retailers

An e-commerce operator monitors 50 products on Amazon, eBay, and Walmart. A twice-daily agent runs Search per product per engine, passing marketplace-specific options through Extra Parameters, and extracts price, seller, and availability from the parsed listings. Changes beyond the margin threshold post to [Slack](/integrations/slack) with the listing linked. When a competitor's price move would trigger a repricing response in the operator's own catalog, that response waits for a merchandiser, because matching a pricing error downward is a mistake you cannot always take back.

Local presence audit for a multi-location business

A franchise operator with 30 locations audits its Google Maps presence monthly. The agent resolves each city with Search Locations to get the canonical location string, runs Search against Google Maps for the service keyword, and records where each location ranks in the local results, and which competitors outrank it. The consolidated report goes to the marketing lead with the locations that fell out of the top results flagged for review, feeding the agency brief with evidence instead of anecdotes.

Human-in-Loop Highlight

Search is the only action in this connector that spends money, and it spends it one query at a time against a finite monthly pool, which is exactly how a mis-scoped loop quietly empties a budget. A keyword list that grows from 400 to 4,000 rows, or a retry branch that re-runs live queries instead of reading the archive, burns quota that cannot be refunded. So FlowRunner agents treat Get Account as a pre-flight check, not a dashboard: before any batch, the agent projects the run's cost against remaining searches, and when the projection crosses the team's budget line, the run pauses and the SEO lead decides: run it whole, trim it, or defer to the reset. The agent also enforces the cheap path by default, reaching for Get Search by ID whenever the question can be answered by a search that already ran. Credits are spent on purpose or not at all.

Agent processes routinely
Detects exception requiring judgment
Clear match Continues automatically
Ambiguous Routes to human via preferred channel
Human decides
Agent resumes with decision

Agent Capabilities

4 actions

Search

2
  • Search Runs a live search against any of 15 engines, including Google Search, Maps, News, Images, Scholar, and Jobs, plus Bing, DuckDuckGo, Yahoo, YouTube, Amazon, eBay, Walmart, Baidu, and Yandex, and returns fully parsed JSON. Common parameters cover query, location, language, country, Google domain, result count, pagination, and device; engine-specific options pass through Extra Parameters. The one action here that consumes a credit.
  • Get Search by ID Retrieves the parsed JSON of a previously executed search from the Search Archive by its search ID, within your plan's retention window. The free re-read that keeps analysis, retries, and audits from re-spending quota.

Account

1
  • Get Account Returns plan name, monthly search quota, searches used, remaining searches and credits, and the hourly rate limit, without consuming a credit. The pre-flight check behind every budgeted batch.

Locations

1
  • Search Locations Looks up the canonical locations SerpApi supports for the Search operation's Location parameter, with canonical name, GPS coordinates, and approximate reach. Free to call, and the difference between "Springfield" and the Springfield you meant.

Frequently Asked Questions

What can FlowRunner do with SerpApi?

FlowRunner agents can run Search, Get Search by ID, and Get Account in SerpApi, plus 1 more action.

Does connecting SerpApi to FlowRunner require OAuth?

No. SerpApi connects to FlowRunner with an API key, no OAuth flow required.

Can SerpApi trigger a FlowRunner workflow automatically?

SerpApi doesn't currently expose triggers in FlowRunner. It connects as an action step inside workflows started by another trigger.

Start building with SerpApi

$100 in credits. No card required. Connect in minutes.