Skip to content
MCP server

WebSearch MCP Server

Zero-API-key Go MCP server for multi-engine web and academic search, PDF parsing and fetch

Runs locallyStatus and history
Install
Pricing
Free
Platforms
CLI, self-hosted, macOS
Runs
Locally
Version
3.10.7
Listing updated
today

Similar MCP servers

About WebSearch MCP Server

WebSearch MCP Server is a free MCP server for web search and scraping. It runs locally on your machine and is started by your AI client.

Written in Go, this server searches without requiring any API key and does its scoring, caching and deduplication locally rather than through a third-party aggregator. The README is in Chinese and English and stresses use from mainland China. It combines general engines such as Baidu, Bing and DuckDuckGo, four hosted zero-key MCP upstreams (Parallel, AnySearch, Exa, Firecrawl) and nine academic engines queried in parallel.

Results are fused with reciprocal rank fusion, relevance boosts and MMR to reduce mirrored reposts. Academic hits are scored on citations, venue and PDF availability and de-duplicated by DOI. cleanfetch reads pages and PDFs in segments with SSRF and DNS-rebinding protection, falling back to Jina Reader. With a token, PDFs can be parsed with MinerU. Optional extras are LLM summaries through an OpenAI-compatible API, a local dashboard and Prometheus metrics, and file_search over an Everything index.

It runs over stdio (npm package or MCPB bundles for macOS, Linux and Windows) or as an HTTP daemon on 127.0.0.1:8338. The README says up to five tools are exposed and that search quality improves with a tailored config.yaml.

Install

Run in a terminal

claude mcp add websearch-mcpserver -- npx -y https://github.com/daidaiJ/websearch-mcpserver/releases/download/v3.10.7-registry/websearch-mcpserver-darwin-amd64.mcpb

Add to claude_desktop_config.json

{
  "mcpServers": {
    "websearch-mcpserver": {
      "command": "npx",
      "args": [
        "-y",
        "https://github.com/daidaiJ/websearch-mcpserver/releases/download/v3.10.7-registry/websearch-mcpserver-darwin-amd64.mcpb"
      ]
    }
  }
}
Install in CursorOpens Cursor, which asks you to confirm. Or add it by hand:

Add to ~/.cursor/mcp.json

{
  "mcpServers": {
    "websearch-mcpserver": {
      "command": "npx",
      "args": [
        "-y",
        "https://github.com/daidaiJ/websearch-mcpserver/releases/download/v3.10.7-registry/websearch-mcpserver-darwin-amd64.mcpb"
      ]
    }
  }
}
Install in VS CodeOpens VS Code, which asks you to confirm. Or add it by hand:

Add to .vscode/mcp.json

{
  "servers": {
    "websearch-mcpserver": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "https://github.com/daidaiJ/websearch-mcpserver/releases/download/v3.10.7-registry/websearch-mcpserver-darwin-amd64.mcpb"
      ]
    }
  }
}

Add to ~/.codeium/windsurf/mcp_config.json

{
  "mcpServers": {
    "websearch-mcpserver": {
      "command": "npx",
      "args": [
        "-y",
        "https://github.com/daidaiJ/websearch-mcpserver/releases/download/v3.10.7-registry/websearch-mcpserver-darwin-amd64.mcpb"
      ]
    }
  }
}

Add to your client's MCP settings (most use this shape)

{
  "mcpServers": {
    "websearch-mcpserver": {
      "command": "npx",
      "args": [
        "-y",
        "https://github.com/daidaiJ/websearch-mcpserver/releases/download/v3.10.7-registry/websearch-mcpserver-darwin-amd64.mcpb"
      ]
    }
  }
}

Using several servers? Build one config file with the MCP config generator.

Tools this server provides (2)

Taken from the project documentation; the running server may differ.

cleanfetch
Reads web pages and PDFs in segments, including saved long results.
file_search
Finds local files through an Everything index.

What it does

Zero-key search
Works without API keys using local engines or hosted MCP upstreams.
Academic search
Nine academic sources queried in parallel with DOI de-duplication.
Result fusion
RRF ranking, boosts and MMR diversity reranking.
Safe page fetch
cleanfetch has SSRF and DNS-rebinding protection.
PDF parsing
MinerU-based parsing when a token is configured, with a text fallback.
Two run modes
stdio client or HTTP daemon with dashboard and metrics.

Good for

  • Give a coding agent free web search from China
  • Search academic papers and rank by citations
  • Read long pages or PDFs in chunks
  • Find local files quickly through Everything

Pricing

Free

Free to run; optional third-party keys

WebSearch MCP Server FAQ

Does it need an API key?

No. Keys are only used when you provide them.

Which clients are named?

Claude Code, Qwen Code and Cursor.

How many tools does it expose?

Up to five.

What language is it written in?

Go.

Is WebSearch MCP Server free?

Yes. WebSearch MCP Server is free to use.

How do I add WebSearch MCP Server to Claude Code?

Run this command in your terminal: claude mcp add websearch-mcpserver -- npx -y https://github.com/daidaiJ/websearch-mcpserver/releases/download/v3.10.7-registry/websearch-mcpserver-darwin-amd64.mcpb

What are the best alternatives to WebSearch MCP Server?

The closest alternatives to WebSearch MCP Server are Scrapling MCP Server, Firecrawl MCP Server, Douyin_TikTok_Download_API, SearXNG Search and machine_internet.

Where this comes from

  • Description: written from the project's README.
  • Install steps: from the official MCP Registry entry.

How AIPoweredBot checks servers