Infomesh

Name: Infomesh
Author: dotnetpower

@dotnetpower

Infomesh について

🕸️ Fully decentralized P2P search engine for LLMs — no API key, no billing, forever free. MCP-native web search via Kademlia DHT + libp2p + FTS5. Production stable ✅ → pip install infomesh

#web-search #search-engine #p2p #decentralized #free #mcp-server #crawler #web-crawler #offline #privacy #llm #peer-to-peer #self-hosted #python #open-source

基本情報

カテゴリ

検索

ライセンス

MIT

ランタイム

python

トランスポート

stdio

公開者

dotnetpower

投稿者

Moonhyuk Choi

設定

以下の設定を使って、このサーバーを MCP 対応クライアントに追加してください。

{
  "mcpServers": {
    "infomesh": {
      "command": "uvx",
      "args": [
        "infomesh",
        "mcp"
      ],
      "env": {}
    }
  }
}

ツール

Search the web via InfoMesh P2P search engine. Returns ranked results with optional full content, RAG chunking, and score explanations. Example: web_search(query='python asyncio tutorial')

Fetch full text of a specific URL. Returns cached content or crawls live. Max 100KB. Example: fetch_page(url='https://...')

Add a URL to the crawl queue and index it. Rate limited to 60/hour. Example: crawl_url(url='https://example.com', depth=1)

Cross-reference a claim against indexed web content. Returns verdict with supporting/contradicting sources. Example: fact_check(claim='Python was created in 1991')

Node status: index size, peer count, credit balance, search quota, and analytics. Example: status()

概要

What is Infomesh?

Infomesh is a fully decentralized peer-to-peer search engine for LLMs, accessible via the Model Context Protocol (MCP). It requires no API keys, credit cards, or usage caps, and is designed for AI developers, open-source assistants, and researchers who need free real-time web search.

How to use Infomesh?

Install Infomesh with pip install infomesh or via the uv package manager. Run the node to join the P2P network, then search from any MCP client, the CLI (infomesh search), Python code, or a local HTTP API. No sign-up or configuration is needed beyond installation.

Key features of Infomesh

Fully decentralized P2P network with no central server
No API key, billing account, or usage cap required
Offline search using a local index
Self-hosted; you own your node and data
Crawl and index your own URLs
Ed25519 cryptographic identity and encrypted transport

Use cases of Infomesh

AI assistants (e.g., Claude, VS Code Copilot) that need free web search via MCP
Local LLMs and open-source projects that cannot afford commercial search APIs
Researchers crawling and indexing custom content for offline use
Developers embedding search in applications via Python or HTTP endpoints

FAQ from Infomesh

How does Infomesh compare to other free web search MCP servers?

Infomesh is the only fully decentralized, offline-capable MCP server with no API key, unlimited free search (credit-based), and support for crawling your own URLs. It does not rely on a central provider or SaaS dependency.

What are the runtime requirements?

Infomesh requires Python 3.12+ and a Linux terminal (Ubuntu, Debian, etc.). No prior Python or developer experience is needed. The uv package manager is optional but recommended.

Where is my data stored?

Search queries are processed locally or routed as hashed keywords through the DHT. Your data never leaves your node unless you choose to share it. Crawled content is fingerprinted, signed, and published to the DHT—no central server stores your search history.

Are there any limits on usage?

Search is credit-based: you earn credits by contributing crawling to the network. There are no per-query costs, rate limits per dollar, or fixed monthly caps. New nodes have a 24-hour probation period, and per-keyword DHT publish rate is limited to 10 per hour per node.

What transports and authentication does Infomesh support?

Infomesh supports MCP (for AI assistants), CLI, Python package, and local HTTP API. Authentication uses Ed25519 cryptographic keys generated on first launch—no API keys or billing accounts are required. All peer-to-peer communication is encrypted via libp2p's Noise protocol.

「検索」の他のコンテンツ

duckduckgo-search MCP Server

zhsama

G-Search MCP

jae-jae

A powerful MCP server for Google search that enables parallel searching with multiple keywords simultaneously.

268

Version 2.0 is here!!

mixelpixx

MCP Server built for use with Claude Code, Claude Desktop, VS Code, Cline - enable google search and ability to follow links and research websites

249

Kagi MCP server

kagisearch

The Official Model Context Protocol (MCP) server for Kagi Search & other tools.

435

Google Search Tool

web-agent-master

A Playwright-based Node.js tool that bypasses search engine anti-scraping mechanisms to execute Google searches. Local alternative to SERP APIs with MCP server integration.

611