Document Weaver

한국어 README

Stop copy-pasting. Drop files, get notes.

Document Weaver converts Word, PDF, PowerPoint, Excel, and HWP files into clean Markdown — instantly. Drag files onto the window, point a watch folder at your Downloads, or pick files from the command palette. Images are extracted automatically. Front matter is injected. Done.

• 📄 DOCX — headings, bold, italic, tables and images, all preserved
• 📑 PDF — text layer extracted with automatic heading detection
• 📊 PPTX — single long-form note or one note per slide, with speaker notes
• 📈 XLSX / XLS — every sheet becomes a clean GitHub-flavored table
• 🇰🇷 HWP / HWPx — Korean office format (beta)
• 📝 TXT / CSV — verbatim or auto-formatted as table

No API keys. No cloud. Fully offline.
The local-file companion to Confluence Weaver.

---

Supported Formats

Format	Extension	Fidelity
Word	.docx	★★★★ — headings, bold/italic, tables, images
PowerPoint	.pptx	★★★☆ — slide titles, bullets, notes, images
PDF	.pdf	★★★☆ — text layer only; scanned PDFs generate a stub note
Excel	.xlsx / .xls	★★★☆ — each sheet becomes a GFM table
HWP	.hwp	★★☆☆ ⚠ beta — binary format, best-effort
HWPx	.hwpx	★★★☆ ⚠ beta — ZIP+XML, better than HWP5 binary
Plain text	.txt / .csv	★★★★ — verbatim / GFM table

---

Installation

BRAT (recommended)

1. Install the BRAT plugin
2. BRAT settings → Add Beta Plugin → enter GS-AX/doc-weaver

Manual

1. Download main.js and manifest.json from Releases
2. Copy both files to .obsidian/plugins/doc-weaver/ inside your Vault
3. Obsidian → Settings → Community Plugins → enable Doc Weaver

---

Usage

Import entry points

Method	How
Command palette	Doc Weaver: Import file… → system file picker (multi-select supported)
Drag & drop	Drop one or more supported files onto the Obsidian window
Watch folder	Configure inbox folders in settings; new files are converted automatically

Output

Converted notes are saved to the configured destination folder (default: Imported/):

---
source_file: "report.docx"
source_format: "docx"
imported_at: "2026-05-23T10:00:00+09:00"
---

# Report Title
...

Embedded images are extracted to Imported/_assets/<note-name>/image-001.png and linked with ![[image-001.png]] (or a markdown link, per setting).

Completion notices

• Single file: ✅ report.docx → Imported/report.md (12 headings, 3 images)
• Bulk: ✅ 5 files imported (1 warning) → Imported/
• Errors are appended to Imported/_import_errors.md

PowerPoint modes

Mode	Output
Single note (default)	One .md file with ## Slide N sections separated by ---
Per-slide	One .md per slide + an index note with [[wikilinks]]

---

Settings

Output

Setting	Default	Description
Destination folder	Imported	Vault folder for converted notes
Asset subfolder	_assets	Sub-path for extracted images
Filename collision	number	skip / overwrite / add number suffix
PowerPoint output	single	Single note or per-slide notes
Use wikilinks	ON	![[...]] vs  for images
Open after import	ON	Open the note after conversion (skipped for bulk)

Watch Folder

Setting	Default	Description
Watch folders	(none)	OS paths to monitor (add multiple)
Watch interval (min)	5	Set to 0 to disable
Watch subfolders	OFF	Recursively watch inbox subfolders
After import	archive	archive / delete / keep
Archive folder	(none)	OS path to move originals after conversion

Advanced

Setting	Default	Description
Show HWP beta features	OFF	Enable HWP/HWPx conversion (limited quality)
Language	Auto	Auto / English / 한국어 / 日本語 / 中文

---

v1.0 Non-Goals

• No reverse export — Markdown → Word/PDF not supported
• No OCR — scanned PDFs produce a stub note only
• No cloud sync — local files only (use Confluence Weaver for remote)
• HWP/HWPx — beta quality; complex formatting and merged table cells may be lost
• PDF tables — rendered as plain text in v1 (heuristic reconstruction planned for v2)

---

License

MIT © GS-AX