Metasearch FAQs

Question 1

What is Metasearch?

Accepted Answer

Metasearch is an open-source backend that aggregates search results from various providers (web, academic, developer, knowledge) into a unified, structured JSON API and MCP server, specifically designed for AI agents and LLM workflows.

Question 2

Which types of search providers does Metasearch support?

Accepted Answer

Metasearch supports a wide range of providers including general web search (DuckDuckGo, Bing), academic sources (arXiv, PubMed), developer resources (GitHub, Stack Overflow), and knowledge bases (Wikipedia), with options for hosted Google APIs (SerpBase, Serper).

Question 3

How does Metasearch benefit AI agents and LLMs?

Accepted Answer

It provides machine-consumable, predictable JSON search results with stable field names, robust error handling, deduplication, and provider-level metadata, enabling agents and LLMs to reliably access and utilize diverse search data without complex parsing.

Question 4

Does Metasearch directly scrape Google search results?

Accepted Answer

No, Metasearch intentionally avoids direct Google scraping due to reliability concerns. Instead, it integrates with reliable hosted Google providers like SerpBase and Serper to offer robust Google search capabilities to AI agents and LLMs.

Question 5

What are the primary integration methods for Metasearch?

Accepted Answer

Metasearch offers an HTTP API with OpenAPI documentation for easy integration into applications and an MCP (Model Context Protocol) server over stdio, making it compatible with agent tooling such as Claude Desktop and Continue.

Metasearch

Metasearch

主な機能

ユースケース

主な機能

ユースケース