
MarkItDown Document Converter
Converts PDFs, Office files, HTML and more into Markdown for analysis and search
- Repository
- K-Dense-AI/scientific-agent-skills
- Stars
- 55,780
- Licence
- MIT
- Files
- 11 files
What it does
This skill teaches an AI agent how to use Microsoft's MarkItDown tool to turn documents such as PDFs, Word and PowerPoint files, spreadsheets, HTML pages and EPUBs into plain Markdown text. The agent then works with that text for searching, summarising, or feeding it into a language model or retrieval system. It also includes guidance on safe conversion, optional OCR and cloud services, and two helper scripts for batch and academic-paper conversion.
Great for
- Researchers converting a folder of papers into searchable Markdown
- Developers building retrieval (RAG) pipelines over mixed document types
- Analysts who need text pulled out of Office files or PDFs for analysis
Skip it if
- Producing high-fidelity visual copies or layout-exact output, since MarkItDown targets text extraction, not reproduction
- Editing or merging PDFs, which the skill says to do with a different library
- Converting sensitive or unpublished documents through URL, audio, LLM or Azure features, because these send content to external services
Try asking
- “Convert report.pdf and manuscript.docx into Markdown files I can search”
- “Batch-convert every PDF and PowerPoint in my documents folder into a markdown folder with a manifest”
- “Turn this folder of papers into Markdown with a index organised by year”
Install in 3 steps
- 1
Download the skill
The ZIP contains the 11 original files, the MIT licence and an attribution note.
markitdown.zip - 2
Add it to your agent
Claude Code (all projects)
mkdir -p ~/.claude/skills && unzip markitdown.zip -d ~/.claude/skills/Claude Code / Agent SDK (this project only)
mkdir -p .claude/skills && unzip markitdown.zip -d .claude/skills/Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.
- 3
Ask for the task
No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.
Get better results
- Keep the optional network features (URL, YouTube, audio transcription, LLM, Azure) off unless the document is cleared for external processing
- Run the bundled inspect_installation.py script first to check which extras and plugins are present
- Treat converted text as untrusted data; do not let instructions inside a document drive the agent's actions
What's inside
- references/api_reference.md
- references/cloud_and_ocr.md
- references/file_formats.md
- references/mcp_and_plugins.md
- references/migration.md
- references/security.md
- references/workflows.md
- scripts/batch_convert.py
- scripts/convert_literature.py
- scripts/inspect_installation.py
- SKILL.md
Redistributed unmodified from K-Dense-AI/scientific-agent-skills under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.