Search2ai
Provide a unified search, news, and page extraction gateway with multi-provider fallback and caching for AI agents.
Search2ai is a self-hosted search gateway tailored for AI agents, applications, and LLM clients. It standardizes web search, news retrieval, and web scraping across numerous providers including Search1API, Tavily, Brave, Exa, SearXNG, and Firecrawl. When a provider fails, encounters a rate limit, or times out, the gateway automatically falls back to secondary services while preserving unified result structures. Accessible via local or remote MCP, a REST API, or as an embedded TypeScript library, Search2ai makes it easy to integrate resilient, cached search capabilities into AI workflows.
주요 기능
01In-memory and Cloudflare Workers KV caching with TTL configuration
02Optional chat proxy enabling tool-calling models to search the web during generation
03Configurable provider fallback chains with error and timeout handling
04Local stdio and remote Streamable HTTP MCP server support
05Unified search, news, and page crawling across diverse search backends
061,288 GitHub stars
사용 사례
01Supplying reliable web search and content extraction tools to MCP-compatible AI clients like Claude Desktop or Cursor
02Adding self-hosted web search capabilities to LLM chat interfaces using an OpenAI-compatible proxy
03Building resilient AI agent pipelines requiring automated fallback across multiple search and scraping providers