Drop any file. Get beautiful Markdown.
Turn PDFs, Word docs, PowerPoints, HTML pages, spreadsheets, images, and code into clean, readable Markdown you'd actually want to open.
┌──────────┐
.pdf .docx .pptx │ │ pretty Markdown
.html .csv .py ──▶│ markitme │──▶ with frontmatter,
.jpg .tsx .json │ │ callouts, and TOC
└──────────┘
Install
Usage
markitme report.pdf
markitme report.pdf --pretty
markitme report.pdf --pretty --obsidian
markitme report.pdf --pretty --stdout | pbcopy
markitme ./docs --human --output ./vault--human is the one-flag shortcut: turns on --pretty, --obsidian, and --wiki (for folders).
What the flags do
| Flag | What it does |
|---|---|
--pretty |
Contextual emoji headings (80+ keywords), info badge, callouts from Note/Warning/Tip/Important patterns, collapsible long quotes, auto-detected code fence languages |
--obsidian |
YAML frontmatter (title, tags, aliases, source, author), image wikilinks (![[file.png|alt]]), cross-document [[links]] in batch mode |
--wiki |
Master README with table of contents, per-file sections, "See Also" cross-links by keyword overlap, typed file index |
--toc |
Injects a table of contents from headings |
--strip |
No decoration at all. Just the raw converted content |
--stdout |
Print to stdout instead of writing a file. Pipe to anything |
--enhance |
TL;DR summary + heading rewrite via local Ollama. No API keys |
--reflect |
Appends a "Key Insights" section generated by Ollama |
--human |
--pretty + --obsidian + --wiki (on folders) |
Before / After
Input: A 15-page architecture PDF
markitme report.pdf (no flags):
## Page 1
Platform Architecture
This document covers our system design.
## Page 2
Components
API Gateway...
markitme report.pdf --pretty --obsidian:
--- title: Report tags: [markitme, pdf, architecture] author: Jane Doe created: 2026-04-11 12:00 ---
# 🏗️ Platform Architecture **Built with MarkitMe** · Source: PDF · 15 pages · Author: Jane Doe · Converted April 11, 2026 --- ## ⚙️ Components > [!note] > All services communicate over gRPC. | Service | Status | |---------|--------| | API | Live | | Auth | Beta |
Folder → Wiki
markitme ./project-docs --human --output ./wiki
Creates:
wiki/
├── README.md ← master doc with TOC, stats, cross-links
├── _index.md ← file listing organized by type
├── architecture.md
├── roadmap.md
├── api-spec.md
├── attachments/
│ ├── diagram.png
│ └── screenshot.jpg
├── summaries/
└── sources/
The master README links everything together with a table of contents, per-file sections, and auto-detected "See Also" connections between documents that share topics.
Document metadata
PDF and Word files have real metadata — author, title, creation date, subject. markitme extracts it and puts it in:
- The YAML frontmatter (
--obsidian) - The badge line (
--pretty) - The
result.metadatadict (library API)
Library API
import markitme md = markitme.to_markdown("report.pdf", pretty=True) result = markitme.convert("slides.pptx", pretty=True, obsidian=True) result.markdown result.title result.images result.page_count result.metadata results = markitme.convert_folder("./docs", output="./vault", pretty=True, wiki=True)
Supported formats
| Category | Formats |
|---|---|
| Documents | PDF, DOCX, PPTX, HTML, CSV, TSV, TXT, MD |
| Images | PNG, JPG, GIF, WebP, BMP, TIFF, SVG |
| Code | 50+ languages — auto-detected syntax highlighting |
Local LLM enhancement
pip install "markitme[enhance]"
markitme paper.pdf --enhance --reflect
Uses whatever model you have in Ollama (prefers llama3.2, gemma2, mistral). Falls back silently if Ollama isn't running. No API keys, no cloud, everything local.
License
MIT