mcp-logseq

Overview

The mcp-logseq server enables Claude to interact directly with LogSeq knowledge graphs via its HTTP API. It provides comprehensive tools for reading, creating, updating, and organizing pages within LogSeq, along with advanced features like semantic vector search and DB-mode graph support. The server implements intelligent markdown parsing to convert structured content into LogSeq's native block format, maintaining hierarchy and metadata.

Try asking AI

After installing, here are 5 things you can ask your AI assistant:

• list_pages

  Browse your LogSeq graph pages

  Try:Show me all my pagesTry:List all pages in my knowledge base
• get_page_content

  Read the content of a specific LogSeq page

  Try:Get my project notesTry:Show me the content of my journal entry
• create_page

  Create a new LogSeq page with structured markdown content

  Try:Create a meeting notes page with agenda itemsTry:Add a new project planning page
• update_page

  Modify an existing LogSeq page with append or replace modes

  Try:Update my task listTry:Add progress updates to my project page
• delete_page

  Remove a page from the LogSeq graph

  Try:Delete the old draft pageTry:Remove the completed project page
• search

  Find content across the LogSeq graph by keywords

  Try:Search for 'productivity tips'Try:Find all references to 'machine learning'
• query

  Execute LogSeq DSL queries to find specific content

  Try:Find all TODO tasks tagged #projectTry:Show me all completed tasks
• find_pages_by_property

  Search pages by their properties (DB-mode only)

  Try:Find all pages with status = activeTry:Get pages tagged as 'important'
• get_page_backlinks

  Find pages that link to a specific page

  Try:What links to this page?Try:Find all pages referencing my project plan
• rename_page

  Rename a page and update all references to it

  Try:Rename 'Old Project Name' to 'New Project Name'Try:Change 'Draft' to 'Final'
• update_block

  Edit block content by its UUID

  Try:Update this specific block textTry:Fix the typo in this paragraph
• vector_search

  Semantic search by meaning (requires vector setup)

  Try:Find notes about shadow work or JungTry:What relates to deep work?

Comparable tools

Installation

{
  "mcpServers": {
    "mcp-logseq": {
      "command": "uv",
      "args": ["run", "--with", "mcp-logseq", "mcp-logseq"],
      "env": {
        "LOGSEQ_API_TOKEN": "your_token_here",
        "LOGSEQ_API_URL": "http://localhost:12315"
      }
    }
  }
}

FAQ

How do I set up vector search with this MCP server?

Vector search is optional and requires Ollama (for embeddings) and LanceDB (vector database). See the [VECTOR_SEARCH.md](VECTOR_SEARCH.md) guide for full setup instructions.

What LogSeq features does this server support?

The server supports page management, block operations, property queries, namespaces, backlinks, DB-mode properties, and optional semantic vector search. It also includes smart markdown parsing to maintain hierarchical structure.

Compare mcp-logseq with