SerpApi

SerpApi is infrastructure that runs real searches against Google and 20+ other engines/marketplaces and returns the results as structured JSON, so a caller doesn't have to build and maintain its own scraper, proxy pool, or CAPTCHA-solving pipeline. Each engine (Google Search, Google Maps, Amazon, YouTube, etc.) is its own dedicated API endpoint with engine-specific parameters. Access is via a REST API authenticated with an api_key query parameter, official client libraries for Python/JS/Ruby/Go/Java/.NET, and a hosted MCP server for AI agents.

• 70+ engine-specific search APIs (Google Search/Images/Maps/Flights/Hotels/Scholar, Bing, DuckDuckGo, Baidu, Yandex, Yahoo, Naver)
• E-commerce search APIs: Amazon, eBay, Walmart, The Home Depot
• Local/vertical APIs: Google Maps, Apple Maps, Yelp, Tripadvisor, OpenTable, Google Flights/Hotels
• YouTube search, video, and transcript APIs
• Real-time rendered results with geolocation targeting and CAPTCHA handling
• Structured JSON output normalized per engine (organic results, local pack, knowledge graph, shopping, ads, news)
• ZeroTrace Mode for privacy-sensitive querying
• Official MCP server exposing a unified multi-engine search tool

Teams building products on top of search-engine or marketplace data - SEO tooling, price/market monitoring, competitive intelligence, or AI agents that need grounded, real-time search results - who want to avoid maintaining their own scraping and proxy infrastructure. The vendor names Nvidia, Shopify, Uber, and Morgan Stanley as customers.

Pricing is metered by monthly search count with a fixed hourly throughput cap per plan, so bursty workloads can hit the throughput ceiling before the monthly quota. Each engine is a separately structured API, so multi-engine users must account for different parameters and responses. This research did not retrieve a self-serve OpenAPI specification, so teams that require a machine-readable contract should verify current availability with SerpAPI.

Setup assessment: 5m

The documented pattern is a single HTTP GET with query parameters (q, location, gl, hl, api_key) and no OAuth flow or webhook configuration required for basic search calls.

Prerequisites

• SerpApi account and API key

Authentication

api_key query parameter (SerpApi private key) on each request

Base URL

https://serpapi.com/search

Machine-readable API

No OpenAPI link in the reviewed source set

• Google Search (engine=google) - organic results, local pack, knowledge graph, shopping, ads, news
• Google Maps / Google Local - local business search results
• Google Flights / Google Hotels - travel search results
• Google Scholar - academic search results
• Amazon / eBay / Walmart / Home Depot search - e-commerce listing data
• YouTube search, video info, and transcript endpoints

• Machine-readable OpenAPI/Swagger specification (not found; docs are per-engine reference pages)
• Exact per-engine rate limits beyond the plan-level searches/hour caps
• GitHub star counts for individual SerpApi repos
• LinkedIn/X follower counts
• A machine-readable OpenAPI specification was not included in the reviewed source set.