Office Docs Generator: Markdown to Word, PDF, PPTX, XLSX

Turn text and data into real office files with one call:

• Markdown or HTML → Word (DOCX) or PDF: headings, bold/italic, links, nested and numbered lists, quotes, tables and code blocks become proper Word styles. PDFs get a title, a primary language and structure-ordered tabs, so they pass more accessibility checks. Right-to-left languages (Arabic, Hebrew, Persian, Urdu...) get right-to-left paragraphs, lists and tables.
• Slides → PowerPoint (PPTX) or PDF: a title slide, bullet slides, speaker notes and native, editable charts (column, bar, line, pie, doughnut, area, scatter), with 8 themes.
• Tables → Excel (XLSX): one or more sheets, styled tables, formulas (with computed values), number formats, frozen header rows and native charts. CSV works too.

The result is a download URL, so an AI agent or an automation (Make, n8n, Zapier, your own code) can hand the file straight to a user.

Why use it

• Made for agents. One flat input, one output row with the file's URL. Works as an MCP tool.
• Real files. Native Word styles, editable PowerPoint charts, Excel formulas with cached results, all in the standard Office Open XML formats (checked with independent readers: python-docx, python-pptx and openpyxl).
• Pay per document. You are charged only when a file is produced. Invalid input fails before any charge.
• Private. Everything runs inside the Actor run; no third-party document API is called.

Input examples

A PDF report from Markdown:

{
  "format": "pdf",
  "fileName": "q3-report",
  "markdown": "# Q3 report\n\nRevenue grew **18%**.\n\n| Region | Q3 |\n|---|---:|\n| North | 1,850 |",
  "paperSize": "A4",
  "language": "en-US"
}

A PowerPoint deck with a chart:

{
  "format": "pptx",
  "title": "Board update",
  "subtitle": "October",
  "theme": "Harbor",
  "slides": [
    {"title": "Agenda", "bullets": ["Results", "Plans"], "notes": "Keep it short."},
    {"title": "Revenue", "chart": {"type": "column", "title": "Revenue by quarter",
      "categories": ["Q1", "Q2", "Q3"], "series": [{"name": "2026", "values": [12, 18, 15]}]}}
  ]
}

An Excel workbook:

{
  "format": "xlsx",
  "sheets": [{
    "name": "Sales",
    "rows": [["Region", "Q1", "Q2", "Total"], ["North", 1200, 1850, "=B2+C2"], ["South", 980, 1430, "=B3+C3"]],
    "numberFormats": [{"range": "B2:D3", "code": "#,##0"}],
    "chart": {"type": "column", "title": "Sales", "range": "A1:C3"}
  }]
}

Output

One dataset row per run (also saved as the OUTPUT record):

{
  "fileName": "q3-report.pdf",
  "format": "pdf",
  "url": "https://api.apify.com/v2/key-value-stores/.../records/q3-report.pdf",
  "contentType": "application/pdf",
  "bytes": 29831,
  "title": "Q3 report",
  "pages": 1
}

Files are kept in the run's key-value store (unnamed stores are deleted after 7 days; download or copy what you need).

Use with AI agents (MCP)

Add this Actor to any MCP client through Apify's MCP server:

https://mcp.apify.com?tools=shatrix/office-docs-generator

The agent sees one tool with the input fields above and gets the file URL back.

Use from code

curl -X POST "https://api.apify.com/v2/acts/shatrix~office-docs-generator/run-sync-get-dataset-items" \
  -H "Authorization: Bearer $APIFY_TOKEN" -H "Content-Type: application/json" \
  -d '{"format": "docx", "markdown": "# Hello\n\nGenerated by the API."}'

Pricing

Pay per event: you pay for each generated document (see the Pricing tab). Runs that fail because of invalid input are not charged. Set a maximum cost per run to cap spending.

Limits

• One document per run; up to 200 slides, 50 sheets and 200,000 cells.
• Images in Markdown or HTML are left out: the document engine doesn't import them yet.
• Markdown links are kept (clickable in DOCX and PDF) but not styled in a link colour.
• Fonts: the image adds Noto (most of the world's scripts) and metric-compatible stand-ins for Calibri, Cambria, Arial, Times New Roman and Courier New to the engines' bundled fonts. Checked: Latin, Greek, Cyrillic, Hebrew, Arabic, Devanagari, Bengali, Tamil, Thai, Georgian, Armenian, Chinese, Japanese and Korean all render.
• Right-to-left layout follows language (for example ar or he-IL) and covers Word and PDF documents from Markdown or HTML; slides and spreadsheets stay left to right.
• Tested with long real-world Markdown: a 320 KB README with 53 tables became a 205-page PDF.
• The document engines are young, actively developed open-source projects; check important documents before sending them on.

Credits

Built on open-source engines from the ArtCraft project (WordCraft, DeckCraft, GridCraft and PdfCraft, MIT OR Apache-2.0). Their license texts ship in the Actor's image under /opt/engines/licenses/. This Actor is not affiliated with or endorsed by the ArtCraft project.