A powerful multi-engine conversion tool that transforms complex documents and images into structured Markdown or JSON with built-in OCR and LaTeX support.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install markitdown-file-converter
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install markitdown-file-converter using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is MarkItDown File Converter?
MarkItDown File Converter is a versatile utility within the Openclaw Skills ecosystem designed to bridge the gap between static documents and programmable text formats. It integrates three heavyweight engines—Pandoc, Microsoft's MarkItDown, and Mammoth—to ensure high-fidelity conversion of Word, Excel, PPT, and PDF files. Whether you are dealing with complex grid tables, embedded emojis, or mathematical formulas, this skill intelligently selects the best backend to preserve document integrity.
Beyond simple text extraction, it features robust OCR capabilities for images and scanned PDFs using RapidOCR and PaddleOCR. It also caters to technical users by automatically converting mathematical expressions into LaTeX and extracting embedded images into organized directories. This makes it an essential tool for developers and researchers who need to ingest diverse data sources into LLMs or documentation sites.
MarkItDown File Converter Use Cases
- Converting legacy Word (.doc/.docx) or PDF manuals into clean Markdown for documentation sites.
- Extracting data from Excel spreadsheets directly into Markdown pipe tables for quick reporting.
- Performing OCR on scanned PDF documents or screenshots to extract text and LaTeX formulas.
- Batch processing entire directories of mixed media files into structured JSON for data analysis.
- Transforming PowerPoint presentations into text-based formats for content repurposing.
How MarkItDown File Converter Works
- The skill performs a dependency check to ensure Pandoc, MarkItDown, and OCR engines are available.
- Upon receiving a file, the format detector identifies the file type (e.g., .xlsx, .pdf, .png).
- An intelligent selector chooses the optimal backend: Pandoc for complex Word docs, MarkItDown for Excel/PPT, or OCR for images.
- During processing, it handles specialized tasks like LaTeX conversion for formulas and image extraction to a local /images folder.
- The content is cleaned and formatted into standard Markdown or structured JSON sections based on document headings.
- The final output is saved to a designated directory, maintaining relative paths for any extracted assets.
MarkItDown File Converter Setup
The skill features an automatic installation mode for Openclaw Skills users. To manually trigger the setup or install all backends at once, use the following commands:
# One-key installation of all backends
python scripts/convert.py --install
# Manual installation of specific components
uv pip install mammoth "markitdown[all]" beautifulsoup4
winget install JohnMacFarlane.Pandoc
winget install poppler
MarkItDown File Converter Data Schema & Taxonomy
The skill organizes output into a structured directory to ensure portability. When converting a file named report.docx, the following structure is created:
| File/Folder | Description |
|---|---|
report_converted/ |
The root output directory. |
report.md |
The primary Markdown output file. |
images/ |
Directory containing all extracted images (image_0.png, etc.). |
report.json |
(Optional) Structured JSON with heading, level, and content keys. |
JSON output follows a schema where the document is split into an array of sections based on Markdown header levels.
MarkItDown File Converter Advanced Features
- Multi-engine fallback: Automatically tries Pandoc → MarkItDown → Mammoth if a primary conversion fails.
- Intelligent OCR Pipeline: Uses PaddleOCR Cloud as primary, falling back to local RapidOCR for text and pix2tex for math.
- Batch Mode: Convert entire folders of heterogeneous files with a single
--batchcommand. - Scanned PDF Support: Converts PDF pages to images first to perform high-accuracy OCR on non-selectable text.
- Math Logic: Automatically identifies and wraps mathematical symbols in $inline$ or $$display$$ LaTeX blocks.
SKILL.md
Loading
Related Openclaw Skills
Plane Project Management
A comprehensive integration allowing AI agents to manage projects, tasks, and documentation through the Plane API.
MeiXin Meeting Assistant
An AI-powered meeting assistant specifically designed for the healthcare and insurance industries to automate full-length transcription and structured minutes generation.
Kuakua Navigator
An intelligent discovery skill that connects users to a massive library of validated psychological assessments, cognitive experiments, and browser-based games.
ALAPI Interface Integration Assistant
An expert integration tool for the ALAPI platform that automates documentation retrieval, code generation, and API testing.
ShopGeni
ShopGeni is a powerful AI-driven shopping assistant that enables visual and natural language product discovery and multi-store price comparisons.
InVideo Alternative AI Video Generator
A cloud-powered AI video creation skill that transforms text scripts and raw footage into polished 1080p MP4 videos.
Featured*
