Skip to content
Back to search

markit

SkillWorks with Claude CodeGitHub

Convert files and URLs to Markdown. Supports PDF, DOCX, PPTX, XLSX, HTML, EPUB, CSV, JSON, GitHub URLs, images, audio, ZIP, and more. Use when you need to extract content from any document format.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/shift-labs-ai--markit/badge.svg)](/listings/shift-labs-ai--markit)

HTML

<a href="/listings/shift-labs-ai--markit"><img src="/api/listings/shift-labs-ai--markit/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

# Documents
markit report.pdf
markit document.docx
markit slides.pptx

# Data
markit data.csv
markit config.json
markit schema.yaml

# Web
markit https://example.com/article
markit https://en.wikipedia.org/wiki/Markdown

# Media
markit photo.jpg                          # EXIF metadata
markit recording.mp3                      # Audio metadata

# Text only, skipping image extraction
markit report.pdf --no-images

# Write to file
markit report.pdf -o report.md

# Pipe it
markit report.pdf | pbcopy
markit data.xlsx -q | napkin create "Imported Data"

---

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

README.md · README · 198 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

This file has changed since it was scanned, so this is not exactly the text that was scanned, and no line is marked. The findings are listed with the safety scan above.

  1. # markit
  2. Convert anything to markdown. PDF, DOCX, PPTX, XLSX, HTML, EPUB, Jupyter, RSS, images, audio, URLs, and more. Works as a CLI and as a library.
  3. ```bash
  4. npm install -g @shiftlabs/markit
  5. ```
  6. ---
  7. ## Quick Start
  8. ```bash
  9. # Documents
  10. markit report.pdf
  11. markit document.docx
  12. markit slides.pptx
  13. # Data
  14. markit data.csv
  15. markit config.json
  16. markit schema.yaml
  17. # Web
  18. markit https://example.com/article
  19. markit https://en.wikipedia.org/wiki/Markdown
  20. # Media
  21. markit photo.jpg # EXIF metadata
  22. markit recording.mp3 # Audio metadata
  23. # Text only, skipping image extraction
  24. markit report.pdf --no-images
  25. # Write to file
  26. markit report.pdf -o report.md
  27. # Pipe it
  28. markit report.pdf | pbcopy
  29. markit data.xlsx -q | napkin create "Imported Data"
  30. ```
  31. ---
  32. ## Supported Formats
  33. | Format | Extensions | How |
  34. |--------|-----------|-----|
  35. | PDF | `.pdf` | Built-in Rust engine |
  36. | Word | `.docx` | Built-in Rust engine; headings, images, and tables |
  37. | PowerPoint | `.pptx` | Built-in Rust engine; slides, notes, and tables |
  38. | Excel | `.xlsx` | Built-in Rust engine; sheets become markdown tables |
  39. | HTML | `.html` `.htm` | Built-in Rust engine; scripts/styles stripped |
  40. | EPUB | `.epub` | Built-in Rust engine; spine-ordered chapters |
  41. | Jupyter | `.ipynb` | Built-in Rust engine; cells and outputs |
  42. | RSS/Atom | `.rss` `.atom` `.xml` | Built-in Rust engine; dated feed items |
  43. | CSV/TSV | `.csv` `.tsv` | Built-in Rust engine; markdown tables |
  44. | JSON | `.json` | Built-in Rust engine; formatted code block |
  45. | YAML | `.yaml` `.yml` | Built-in Rust engine; code block |
  46. | XML/SVG | `.xml` `.svg` | Built-in Rust engine; code block |
  47. | Images | `.jpg` `.png` `.gif` `.webp` | Built-in Rust metadata extraction |
  48. | Audio | `.mp3` `.wav` `.m4a` `.flac` | Built-in Rust metadata extraction |
  49. | ZIP | `.zip` | Built-in Rust recursive conversion |
  50. | URLs | `http://` `https://` | Rust fetcher with markdown negotiation |
  51. | Wikipedia | `*.wikipedia.org` | Built-in Rust main-content extraction |
  52. | Code | `.py` `.ts` `.go` `.rs` ... | Built-in Rust fenced code block |
  53. | Plain text | `.txt` `.md` `.rst` `.log` `.vtt` `.srt` | Built-in Rust pass-through |
  54. ---
  55. ## Benchmarks
  56. Non-OCR PDF to Markdown: parsers that read the text layer already inside a
  57. PDF, with no OCR and no model calls. Measured on one machine, with image
  58. extraction off for every parser.
  59. | | markit | liteparse | anydoc |
  60. |---|---:|---:|---:|
  61. | olmOCR-bench score | **45.0%** | 38.8% | 31.2% |
  62. | shitty-pdf-bench text recovered | **100.0%** | 83.2% | 91.0% |
  63. | shitty-pdf-bench conversion time | **23.2s** | 55.4s | 322.2s |
  64. olmOCR-bench is AllenAI's public benchmark. shitty-pdf-bench is ours: 40
  65. hash-pinned public PDFs, 50,884 pages of semiconductor manuals, standards,
  66. RTL documents, scans, and malformed files.
  67. Against the wider non-OCR field, on a 16-document subset covering 3,476
  68. pages and 52 sampled passages:
  69. | | markit | liteparse | anydoc | pypdf | pdfplumber | markitdown | pymupdf4llm |
  70. |---|---:|---:|---:|---:|---:|---:|---:|
  71. | Text recovered | **100.0%** | 98.1% | 94.2% | 84.6% | 73.1% | 73.1% | 90.4% |
  72. | Pages/s | **3,612** | 486 | 247 | 70 | 31 | 25 | 11 |
  73. Method, provenance, and per-category results are in
  74. [rust/BENCHMARKS.md](rust/BENCHMARKS.md); the harness that reproduces them
  75. is in [benchmark/](benchmark/).
  76. ---
  77. ## For Agents
  78. Every command supports `--json`. Raw markdown with `-q`.
  79. ```bash
  80. markit report.pdf --json # Structured output for parsing
  81. markit report.pdf -q # Raw markdown, nothing else
  82. markit onboard # Add instructions to CLAUDE.md
  83. gh skill install shift-labs-ai/markit markit # Install the agent skill
  84. ```
  85. ---
  86. ## Page Markers
  87. Every PDF page opens with a marker, so any line can be traced back to
  88. the page it came from:
  89. ```
  90. <!-- page 12 -->
  91. ```
  92. The number is the 1-based **physical** page — page 12 of the file, which
  93. is not always the number printed on it. Every page gets one, including
  94. blank and image-only pages, so the markers run 1..N with no gaps. They
  95. are HTML comments, so they disappear when the markdown is rendered.
  96. ---
  97. ## SDK
  98. markit is also a Node library. Every format uses the same bundled Rust engine as the CLI on supported macOS and Linux platforms—there is no second fallback implementation:
  99. ```typescript
  100. import { Markit } from "@shiftlabs/markit";
  101. const markit = new Markit();
  102. const { markdown } = await markit.convertFile("report.pdf");
  103. const { markdown } = await markit.convertUrl("https://example.com");
  104. const { markdown } = await markit.convert(buffer, { extension: ".docx" });
  105. ```
  106. ---
  107. ## CLI Reference
  108. ```bash
  109. markit <source> # Convert file or URL
  110. markit <source> -o output.md # Write to file
  111. markit <source> --json # JSON output
  112. markit <source> -q # Raw markdown only
  113. markit <source> --no-images # Skip image extraction
  114. markit <source> -i ./images # Extract images to a directory
  115. cat file.pdf | markit - # Read from stdin
  116. markit formats # List supported formats
  117. markit onboard # Add to CLAUDE.md
  118. ```
  119. ---
  120. ## Development
  121. ```bash
  122. bun install
  123. bun run build:native
  124. bun run dev -- report.pdf
  125. bun test
  126. bun run check
  127. cd rust && cargo test
  128. ```
  129. ## Distribution
  130. Releases publish `@shiftlabs/markit` plus six scoped native binary packages
  131. for macOS and Linux (x64/ARM64, glibc/musl). npm selects the matching binary
  132. through `optionalDependencies`. The former `markit-ai` package is a deprecated
  133. compatibility shim that re-exports this SDK.
  134. ```bash
  135. # 1. Update the same version in package.json, rust/Cargo.toml,
  136. # and npm/*/package.json.
  137. # 2. Verify locally.
  138. bun run verify
  139. # 3. Tag and push. GitHub Actions builds and publishes all seven packages.
  140. git tag v0.6.0
  141. git push origin v0.6.0
  142. ```
  143. The release workflow requires an npm automation token in the repository secret
  144. `NPM_TOKEN`. It publishes platform packages first through `napi prepublish`,
  145. then `@shiftlabs/markit`, then the deprecated `markit-ai` shim, and finally
  146. creates the GitHub release.
  147. ## License
  148. MIT

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.