title: "@stnd/ingest" aliases: [] created: 2026-07-05 07:40 modified: 2026-07-05 19:22 last_audited: 2026-07-14 audit_interval_days: 90 next_audit: 2026-10-12 audit_priority: 3 maturity: sprout mode: read publish: false status: active tags:

• package
• stnd theme: kernel type: package visibility: private

@stnd/ingest

Document-to-Markdown converter for the Standard Framework.

Converts DOCX, PDF, HTML, and RTF files into clean Markdown — ready for @stnd/press to render beautifully.

ELI5

Someone drops a Word doc, a PDF, or a random HTML export into your app — this turns it into plain Markdown so @stnd/press can typeset it properly. The heavy converters (mammoth for Word, pdfjs-dist for PDF) only load when that specific format is actually used.

Use it:

import { convertDocumentToMarkdown } from "@stnd/ingest";
const markdown = await convertDocumentToMarkdown(file);

Philosophy

Strip everything that isn’t content. No formatting noise. The Press will apply beautiful typography automatically.

Supported Formats

| Format | Extensions | Engine | | :-------------------- | :-------------- | :----------------------------------------------------------------------- | | Plain text / Markdown | .txt, .md | Pass-through | | Microsoft Word | .docx, .doc | mammoth (lazy-loaded) | | HTML | .html, .htm | html-to-md | | PDF | .pdf | pdfjs-dist (lazy-loaded) | | Rich Text Format | .rtf | Basic text extraction |

Heavy dependencies (mammoth, pdfjs-dist) are lazy-loaded — they only get imported when you actually convert that format.

Usage

import {
  convertDocumentToMarkdown,
  extractDocumentTitle,
  isSupportedDocument,
} from "@stnd/ingest";

// Check if a file can be converted
if (isSupportedDocument(file)) {
  // Convert to clean markdown
  const markdown = await convertDocumentToMarkdown(file);

  // Extract a title from the resulting markdown
  const title = extractDocumentTitle(markdown, file.name);
}

API

convertDocumentToMarkdown(file): Promise<string>

Accepts a File object (or any object with name, type, text(), and arrayBuffer() methods). Returns clean Markdown.

extractDocumentTitle(markdown, filename?): string

Extracts a title from converted Markdown content. Checks headings first, then the first line, then falls back to the filename.

isSupportedDocument(file): boolean

Checks whether a file can be converted based on its MIME type or extension.

getSupportedMimeTypes(): string[]

Returns all MIME types that can be converted.

getSupportedExtensions(): string[]

Returns all file extensions that can be converted.

Relationship to other packages

User drops a file
       ↓
  @stnd/ingest     →  "Here's clean Markdown"
       ↓
  @stnd/press      →  "Here's beautiful HTML"

@stnd/ingest handles the messy input. @stnd/press handles the beautiful output. Clean separation of concerns.

Notes / Observations

(jot down anything noticed here — quirks, gotchas, ideas)

Todo

• [ ] Maybe relocate to stnd.gd (app need > framework need). [priority:: 3] [token_scale:: 3] [created:: 2026-07-14] [area:: framework]