Serper (serper.dev)
A Google Search results API: you send a query and get Google results back as JSON. It is a SERP API, not an index of its own.
What it offers
- Endpoints: Search, Images, News, Maps, Places, Videos, Shopping, Scholar, Patents and Autocomplete; the website lists no scrape endpoint
- Speed: the website claims results in 1 to 2 seconds
- Billing model: credit-based, with 2,500 free queries and no card required; the website shows no tier names or per-query rates on its front page, so see the pricing page for plans
- Used for: feeding Google results into agents and RAG pipelines, often through agent frameworks that ship a Serper tool
How it differs from the AI-native search APIs
| Serper | Exa, Parallel, Tavily | |
|---|---|---|
| Source of results | Google’s results page | Own index or own retrieval stack |
| Returns | Ranked links, snippets, knowledge-graph and related data | Ranked links plus cleaned page content and excerpts, sometimes synthesized answers |
| Page content | Not on the front page; fetch separately | Built in |
For other SERP-based options see serpapi and the Bright Data SERP API in brightdata.
Limits and legal context
- The website names no operating company, address or country; the privacy policy says only “Serper” is the data controller and that it operates globally. The policy says it logs API requests, including the query and IP address.
- Google’s terms forbid automated access to its results. In December 2025 Google sued serpapi under the DMCA over bypassing its SearchGuard protection; that case does not name Serper, and no ruling about Serper was found. Teams with compliance needs should read Serper’s terms and consider search APIs that run their own index.