Skip to content
MCP server

ResearchZosho

Local deep research assistant that checks each citation and keeps what it finds

Runs locallyStatus and history
Pricing
Open source
Platforms
self-hosted, CLI, Web
Runs
Locally
Version
0.7.7
Listing updated
3 days ago

About ResearchZosho

ResearchZosho is an open-source MCP server for knowledge and memory. It runs locally on your machine and is started by your AI client.

Where most research tools end with a report, ResearchZosho keeps a library. Each fact is noted with its source and a quote, and nothing counts as valid until you accept it. Accepted claims are reused on the next question about the same subject, and the library flags it when a paper you used gets a new version or two of your findings contradict each other.

A run reads in parallel, in more than one language when the question needs it, and writes up sections ending in a table of the facts used. Each cited sentence is checked against its source before it is shelved. One source stays a draft until a second independent one backs it, a retracted paper marks anything citing it as disputed, and you can ban or trust sites.

It runs on your computer with a local model through llama.cpp, Ollama or LM Studio, or with a hosted API. Besides web and paper search it can take your own PDFs and folders, survey a repository or paper, query a read-only database (SQLite, PostgreSQL, MySQL, SQL Server, MongoDB or CSV files), and recommend books from library catalogues.

Install

Run in a terminal

claude mcp add researchzosho -- npx -y @wyrdsekai/researchzosho-mcp

Add to claude_desktop_config.json

{
  "mcpServers": {
    "researchzosho": {
      "command": "npx",
      "args": [
        "-y",
        "@wyrdsekai/researchzosho-mcp"
      ]
    }
  }
}
Install in CursorOpens Cursor, which asks you to confirm. Or add it by hand:

Add to ~/.cursor/mcp.json

{
  "mcpServers": {
    "researchzosho": {
      "command": "npx",
      "args": [
        "-y",
        "@wyrdsekai/researchzosho-mcp"
      ]
    }
  }
}
Install in VS CodeOpens VS Code, which asks you to confirm. Or add it by hand:

Add to .vscode/mcp.json

{
  "servers": {
    "researchzosho": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "@wyrdsekai/researchzosho-mcp"
      ]
    }
  }
}

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

{
  "mcpServers": {
    "researchzosho": {
      "command": "npx",
      "args": [
        "-y",
        "@wyrdsekai/researchzosho-mcp"
      ]
    }
  }
}

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

{
  "mcpServers": {
    "researchzosho": {
      "command": "npx",
      "args": [
        "-y",
        "@wyrdsekai/researchzosho-mcp"
      ]
    }
  }
}

Configuration

  • RESEARCHZOSHO_LIBRARYThe library folder (default ~/researchzosho-library)

What it does

Checked citations
Each cited sentence is compared with its source before it goes on the shelf.
Persistent library
Findings are kept so later questions on the same topic build on them.
Claim review
Accept, dispute or retire each claim; the library uses what you accepted.
Own documents
Shelve PDFs, web pages or whole folders, plus chats, reading lists and drafts.
Database access
Give the library read-only access to a database and have reports cite the query run.
Local or hosted models
Use llama.cpp, Ollama, LM Studio or a hosted API.
Repository survey
Read a code repository, paper or site and list research questions about it.

Good for

  • Check what studies say about a treatment before a vet or doctor visit
  • Trace family records across sources in several languages
  • Compare forty saved papers for agreement and gaps in a thesis chapter
  • Ask questions of your own database and get a sourced write-up
  • Find which book to read first on a subject

Pricing

Open source

Free and open source; model costs depend on the model you choose

Prices change. Check researchzosho.org before you buy.

ResearchZosho FAQ

Does it run locally?

Yes. It runs on your own computer with a local LLM or a hosted API, and the library folder defaults to ~/researchzosho-library.

Is it open source?

Yes, the project describes itself as free and open source.

How are bad sources handled?

A single source is only a draft until an independent second source backs it, retracted papers dispute what cites them, and banned sites are not used.

How do I install the MCP server?

It is published on npm as @wyrdsekai/researchzosho-mcp and as an OCI image at ghcr.io/wyrdsekai/researchzosho.

Is ResearchZosho free?

Yes. ResearchZosho is open source, so you can use and self-host it for free.

How do I add ResearchZosho to Claude Code?

Run this command in your terminal: claude mcp add researchzosho -- npx -y @wyrdsekai/researchzosho-mcp

What are the best alternatives to ResearchZosho?

The closest alternatives to ResearchZosho are Scrapling MCP Server, Codebase Memory, XERJ, Firecrawl MCP Server and Claude Code Ultimate Guide.

Alternatives to ResearchZosho

  1. MCP server for AI clients to scrape sites with HTTP and stealth browser sessions

    Runs locally
    • 17 tools
    • Open source
    • 104K GitHub stars
    • Web search and scraping
  2. Codebase Memory

    MCP server

    Local MCP server that builds a code knowledge graph for AI coding agents

    Runs locally
    • Open source
    • 48.7K GitHub stars
    • Knowledge and memory
  3. XERJ

    MCP server

    Local search engine for AI agents with cited answers over your code and docs

    Runs locally
    • Open source
    • 19.2K GitHub stars
    • Knowledge and memory
  4. MCP server that lets AI agents search, scrape and interact with web pages

    • 3 tools
    • Remote
    • Freemium
    • 7.8K GitHub stars
    • Web search and scraping
  5. MCP server for searching the Claude Code Ultimate Guide from any MCP client

    Runs locally
    • Open source
    • 6.5K GitHub stars
    • AI models and LLM tooling
  6. SearXNG Search

    MCP server

    MCP server for web search and page reading through a SearXNG instance

    Runs locally
    • 3 tools
    • Open source
    • 1.4K GitHub stars
    • Web search and scraping