infomesh

Infomesh MCP Server

InfoMesh brings peer-to-peer web search to AI assistants through MCP, with a local index that can also be searched offline. Developers and researchers can retrieve pages, queue URLs for indexing, check claims against indexed sources, and inspect their node.

Server facts

infomesh MCP Server

Project

License
MIT

Tech

Runtime / language
python
Transports
stdio
Remote available
No
Authentication
Not specified

Activity

GitHub stars
7
Last commit
Data as of

Tags

  • Python
  • Self Hosted
  • Web Search
  • crawler
  • decentralized
  • free
  • llm
  • mcp server
  • offline
  • open source
  • p2p
  • peer to peer
  • privacy
  • search engine
  • web crawler

Overview

InfoMesh organizes web discovery around participant-run crawling and search nodes rather than a central search service. For AI assistants and developers, its web_search tool retrieves ranked matches and can include page content, chunks for retrieval-augmented generation, and explanations of result scores. A targeted lookup can continue through fetch_page, which supplies text from a cached page or a live crawl. To add material to the searchable collection, crawl_url places a chosen address in the crawling queue for indexing. Local keyword retrieval uses SQLite FTS5 with BM25, while semantic search is optional through ChromaDB.

The workflow also includes fact_check, which compares a claim with indexed material and returns a verdict alongside sources that support or contradict it. The status tool exposes operational details such as connected peers, index size, available credits, search quota, and analytics. Network participation follows a contribution-based credit model, and the local index remains searchable without a network connection. Written in Python and available under the MIT license, InfoMesh connects to MCP clients over stdio without requiring an API key.

infomesh setup notes

  • infomesh records stdio only; confirm a network transport upstream before planning remote access.

Related MCP servers