AnySearch MCP Server offers AI agents a unified, real-time search capability across the general web and 23 specialized domains. It supports parallel batch queries and full-page URL content extraction, providing developers with a versatile tool for information retrieval, with flexible transport options and optional anonymous access.

How It Works

This server functions as a unified skill for AI agents, enabling general web search via natural language, structured queries across 23 vertical domains (finance, academic, security, etc.), and parallel execution of up to five independent queries. It also extracts full web page content into Markdown. The system natively supports Streamable HTTP transport, with SSE and stdio clients connecting via proxy tools like mcp-remote or supergateway.

Quick Start & Requirements

Configuration involves directing AI agents to https://api.anysearch.com/mcp. Streamable HTTP is recommended, requiring direct setup in agent-specific files (e.g., ~/.opencode/config.json). SSE/stdio transports necessitate proxy tools (mcp-remote, supergateway), which require npx and Node.js. An API key from anysearch.com/console/api-keys is optional for higher rate limits; anonymous access is available with reduced limits.

Highlighted Details

• Unified search across general web and 23 vertical domains (finance, academic, security, legal, code, etc.).
• Parallel batch search for up to 5 independent queries.
• Full-page URL content extraction to Markdown (truncated at 50,000 characters).
• Anonymous access supported with lower rate limits.

Maintenance & Community

The provided README does not detail maintainers, community channels (Discord/Slack), sponsorships, or a public roadmap.

Licensing & Compatibility

The README omits explicit licensing information. Consequently, its compatibility for commercial use or integration with closed-source projects remains undetermined.

Limitations & Caveats

Sensitive data should not be used in queries as all requests are sent to https://api.anysearch.com. URL content extraction is limited to HTML pages and truncates at 50,000 characters. SSE/stdio transports require running separate proxy tools, adding setup complexity.