Prima Scholar Library MCP Server

A Model Context Protocol server that provides a local SQLite document library with FTS5 full-text search. Import PDFs, Word documents, Markdown, and plain text files, then search across their content instantly.

Installation

cd prima-scholar-library-mcp
npm install
npm run build

Configuration

Claude Desktop

Add to ~/Library/Application Support/Claude/claude_desktop_config.json:

{
  "mcpServers": {
    "prima-scholar-library": {
      "command": "node",
      "args": ["/path/to/prima-scholar-library-mcp/build/index.js"],
      "env": {
        "RESEARCH_LIBRARY_PATH": "~/.research-library/library.db"
      }
    }
  }
}

Claude Code

Add to .claude.json or project settings:

{
  "mcpServers": {
    "prima-scholar-library": {
      "command": "node",
      "args": ["/path/to/prima-scholar-library-mcp/build/index.js"],
      "env": {
        "RESEARCH_LIBRARY_PATH": "~/.research-library/library.db"
      }
    }
  }
}

Environment Variables

| Variable | Default | Description | |----------|---------|-------------| | RESEARCH_LIBRARY_PATH | ~/.research-library/library.db | Path to the SQLite database file. Created automatically if it does not exist. |

Available Tools

| Tool | Description | |------|-------------| | library_import | Import a local file (PDF, DOCX, Markdown, plain text) into the library | | library_import_from_search | Add a document from bibliographic metadata (no local file needed) | | library_search | Full-text search across all documents with ranked results and snippets | | library_get_document | Get full details of a document by ID, including tags | | library_list | List documents with optional filtering by collection, tag, year, or file type | | library_update | Update metadata fields on an existing document | | library_delete | Permanently delete a document from the library | | library_create_collection | Create a named collection for organising documents | | library_list_collections | List all collections with document counts | | library_add_to_collection | Add a document to a collection | | library_tag | Add or remove a tag on a document | | library_stats | Get aggregate library statistics |

Database

The server uses SQLite with FTS5 for full-text search. The database is created automatically on first run and contains:

• documents — Core document table with title, authors, abstract, content, year, DOI, file path, and notes
• documents_fts — FTS5 virtual table indexing title, authors, abstract, content, and notes
• collections — Named groups for organising documents
• tags — Labels that can be applied to any document
• document_collections / document_tags — Many-to-many join tables

FTS triggers keep the search index synchronised automatically on insert, update, and delete.

Supported File Types

| Extension | Importer | Notes | |-----------|----------|-------| | .pdf | pdf-parse | Extracts text and metadata (title, author, subject, keywords) | | .docx | mammoth | Extracts raw text content | | .md | Built-in | Extracts title from first heading | | .txt | Built-in | Uses filename as title |