Markdown Editor MCP Server
Provides semantic editing tools for Markdown files, allowing structured manipulation of document elements through hierarchical paths rather than raw text operations. Supports navigation, search, content replacement, element insertion/deletion, undo functionality, and YAML frontmatter management.
README
Markdown Editor MCP Server
MCP server providing tools for structured, semantic editing of Markdown files. Unlike standard text editors, this server understands the logical structure of your documents.
Fully compliant with MCP 2025 Standard - includes Tool Search, Examples, Output Schemas, and Dynamic Capabilities.
Installation
From PyPI (recommended)
pip install markdown-editor-mcp-server
From source
git clone https://github.com/KazKozDev/markdown-editor-mcp-server.git
cd markdown-editor-mcp-server
pip install -e .
Claude Desktop Configuration
Add to ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):
{
"mcpServers": {
"markdown-editor": {
"command": "markdown-editor-mcp-server"
}
}
}
Available Tools
Discovery & Scaling
<details> <summary><b>search_tools</b> - Find the right tool</summary>
Why LLMs need this: With 15 tools available, finding the right one can be challenging. This meta-tool helps discover which tool to use for a specific task.
What it does: Searches through all available tools and returns the most relevant ones based on your query.
Parameters:
query(required): Description of what you want to do
Example:
{"query": "replace text"}
</details>
Semantic Editing
<details> <summary><b>get_document_structure</b> - Get document tree</summary>
Why LLMs need this: Large Markdown files are hard to navigate. This tool converts the file into a structural tree, allowing the LLM to understand the hierarchy of headings, paragraphs, and lists.
What it does: Parses the Markdown file and returns a JSON tree of all elements with their semantic paths.
Parameters:
file_path(required): Path to the .md filedepth: Max depth for the tree (default: 2)
Example:
{"file_path": "docs/report.md", "depth": 3}
</details>
<details> <summary><b>search_text</b> - Semantic search and navigation</summary>
Why LLMs need this: Instead of scrolling through an entire file, the LLM can search for specific keywords and get exact semantic paths (e.g., Project > Deadlines > paragraph 2).
What it does: Searches the document for text and returns paths to the containing elements.
Parameters:
file_path(required): Path to the filequery(required): Text to search for
Example:
{"file_path": "todo.md", "query": "urgent"}
</details>
<details> <summary><b>read_element</b> - Read specific element</summary>
Why LLMs need this: Before editing an element, you often need to see its full content. This tool fetches the complete content of a specific block.
What it does: Returns the full content of an element identified by its path.
Parameters:
file_path(required): Path to the filepath(required): Semantic path to the element
Example:
{"file_path": "notes.md", "path": "Features > list 1"}
</details>
<details> <summary><b>replace_content</b> - Precise content replacement</summary>
Why LLMs need this: Overwriting entire files is risky and token-expensive. This tool allows the LLM to replace the content of a specific semantic block without affecting the rest of the document.
What it does: Replaces text in a specific element identified by its path.
Parameters:
file_path(required): Path to the filepath(required): Semantic path (e.g., "Intro > paragraph 1")new_content(required): New text for the block
Example:
{"file_path": "readme.md", "path": "Installation > paragraph 1", "new_content": "Just run pip install."}
</details>
<details> <summary><b>insert_element</b> - Add new content</summary>
Why LLMs need this: Adding new paragraphs, headings, or lists requires understanding document structure. This tool inserts new elements at the right location.
What it does: Inserts a new block (heading, paragraph, list, code block, or blockquote) before or after an existing element.
Parameters:
file_path(required): Path to the filepath(required): Reference element pathelement_type(required): Type (heading, paragraph, list, code_block, blockquote)content(required): Content of the new elementwhere: "before" or "after" (default: "after")heading_level: Level for headings (default: 1)
Example:
{"file_path": "doc.md", "path": "Introduction", "element_type": "paragraph", "content": "New paragraph here."}
</details>
<details> <summary><b>delete_element</b> - Remove content</summary>
Why LLMs need this: Removing specific blocks without affecting surrounding content requires precision. This tool deletes elements by their semantic path.
What it does: Removes a block from the document.
Parameters:
file_path(required): Path to the filepath(required): Semantic path to the element to delete
Example:
{"file_path": "draft.md", "path": "Old Section"}
</details>
<details> <summary><b>move_element</b> - Structural refactoring</summary>
Why LLMs need this: Reorganizing a document is complex. This tool lets the LLM move entire sections (including all nested sub-elements) to a new location with a single command.
What it does: Moves a block of text from source path to target path.
Parameters:
file_path(required): Path to the filesource_path(required): Path of element to movetarget_path(required): Reference path for new locationwhere: "before" or "after" the target (default: "after")
Example:
{"file_path": "draft.md", "source_path": "Contacts", "target_path": "Conclusion"}
</details>
<details> <summary><b>get_context</b> - Context-aware editing</summary>
Why LLMs need this: When editing a single block, LLMs might lose the narrative flow. This tool provides the target element along with its immediate neighbors (before and after).
What it does: Fetches an element and snippets of surrounding blocks.
Parameters:
file_path(required): Path to the filepath(required): Path to the target element
Example:
{"file_path": "article.md", "path": "Body > paragraph 5"}
</details>
<details> <summary><b>update_metadata</b> - YAML Frontmatter management</summary>
Why LLMs need this: Many Markdown tools (Obsidian, Jekyll) use YAML metadata at the top. This tool allows the LLM to manage tags, dates, and properties without messing with the body text.
What it does: Updates or adds YAML Frontmatter to the document.
Parameters:
file_path(required): Path to the filemetadata(required): Dictionary of metadata items
Example:
{"file_path": "post.md", "metadata": {"status": "published", "tags": ["mcp", "ai"]}}
</details>
<details> <summary><b>undo</b> - Revert changes</summary>
Why LLMs need this: Mistakes happen. This tool allows reverting recent operations without manual file restoration.
What it does: Reverts the last N operations on the document.
Parameters:
file_path(required): Path to the filecount: Number of operations to undo (default: 1)
Example:
{"file_path": "document.md", "count": 2}
</details>
File Operations
<details> <summary><b>list_directory</b> - Explore workspace</summary>
What it does: Lists files and folders in a directory, helping the LLM explore the file structure.
Parameters:
path: Directory path (default: ".") </details>
<details> <summary><b>create_file</b> - Create new document</summary>
What it does: Creates a new file with optional initial content.
Parameters:
path(required): Path to the new filecontent: Initial content (default: "")
Example:
{"path": "./notes/todo.md", "content": "# TODO\n\n- Task 1"}
</details>
<details> <summary><b>create_directory</b> - Create folder</summary>
What it does: Creates a new directory.
Parameters:
path(required): Path to the new directory
Example:
{"path": "./docs/archive"}
</details>
<details> <summary><b>delete_item</b> - Cleanup</summary>
What it does: Deletes a file or directory. </details>
Development
# Install for development
pip install -e .
# Run tests
pytest tests/
MCP 2025 Standard Features
This server implements all 2025 MCP improvements:
- Tool Search Tool - Efficiently find the right tool among many options
- Tool Use Examples - Input parameter examples help LLMs use tools correctly
- Output Schemas - Structured output definitions for better client integration
- Dynamic Capabilities - Support for
tools/list_changednotifications
If you like this project, please give it a star ⭐
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。