Markdown Formatter

What the markdown-formatter MCP server does

The markdown-formatter MCP server exposes document conversion, Markdown cleanup, inspection, and batch-processing operations through MCP. It is designed for workflows where Markdown produced by an AI or copied from another source must become a file, a platform-specific message, or structured data.

Supported document outputs include plain text, HTML, PDF, DOCX, LaTeX, RTF, CSV, JSON, XML, XLSX, PNG, and Markdown. HTML output can include GitHub-Flavored Markdown tables and KaTeX math. PDF and image conversion use headless Chromium, while DOCX output preserves common document structures such as headings, lists, and code formatting. Markdown tables can be exported to CSV, or placed into separate XLSX sheets.

Platform converters target Slack, Discord, JIRA, Confluence, AsciiDoc, reStructuredText, MediaWiki, BBCode, Textile, and Emacs Org Mode. An HTML import tool converts a complete document or fragment back to Markdown.

How it works

An MCP client calls an individual tool with Markdown or HTML input and conversion options where supported. The server includes tools for normalizing Markdown, repairing common syntax damage, and reporting lint findings as structured JSON. Repair targets include unclosed fences, malformed tables, stray emphasis markers, missing spaces after headings, and broken links.

Analysis tools return code blocks, links and images, heading-based tables of contents, document counts, estimated reading time, and an overall structural summary. The batch_convert tool accepts multiple documents and multiple target formats in one request. Errors are isolated for each document-format combination, and an optional output directory can be used for file output.

Six built-in prompts combine these operations for common sequences: general conversion, table extraction, PDF and HTML preparation, lint-repair-lint workflows, platform conversion, and document overviews. Two resources expose a supported-format list and a format-selection guide.

Setup and configuration

The documented quick-start method runs the package with npx:

{
  "mcpServers": {
    "ai-answer-copier": {
      "command": "npx",
      "args": ["-y", "@xjtlumedia/markdown-mcp-server"]
    }
  }
}

The package can also be installed globally with npm, after which the markdown-mcp-server command is used in the MCP client configuration. The project provides an online web version, but the MCP setup described here uses the npm package. No API key or other credential is specified in the provided setup instructions.

Tools and capabilities

The 34 listed tools cover five main areas:

• Conversion to files or strings, including HTML, PDF, DOCX, LaTeX, RTF, CSV, JSON, XML, XLSX, PNG, TXT, and Markdown.
• Conversion to collaboration, documentation, forum, and markup formats.
• Import from HTML into Markdown.
• Markdown repair, linting, code and link extraction, table-of-contents generation, and document analysis.
• Multi-document, multi-format batch conversion.

The server also provides generate_html for self-contained HTML returned without file I/O and convert_to_email_html for HTML with inlined styles intended for common email clients.

Limitations and notes

The provided material documents Markdown and HTML conversion behavior, but does not specify every input option, output path convention, operating-system requirement, or client-specific configuration. PDF and PNG generation depend on headless Chromium according to the tool descriptions. Platform output is adapted to each target syntax; it is not necessarily interchangeable with standard Markdown. The documentation lists both 22-format batch conversion language and a broader set of individual tools, so the exact format count depends on whether platform and import capabilities are included.