PandocPro

🚀 PandocPro

AI-powered long-document editing — private, fast, and built for real workflows.

PandocPro is a local-first writing and editing tool that lets you convert, improve, and manage large Word and Markdown documents without token limits, formatting issues, or privacy risks.

Work section-by-section. Use local LLMs. Keep everything on your device.
Perfect for writers, agencies, researchers, and anyone producing long or sensitive documents.

---

🌟 Why PandocPro?

Most AI tools collapse when you give them a 40-page brief, a client report, or a technical document. PandocPro solves this by creating a clean, predictable, private document workflow around AI.

You’ll download PandocPro because you want:

• AI assistance without exposing private files
• Long-document editing without token or context limits
• Clean DOCX ↔ Markdown transitions
• Precise section-based editing and improvements
• A workflow you can trust, not a hacky workaround
• A desktop app (with a .dmg) you can install and run instantly

PandocPro is built for people who write or edit professionally — not for quick notes.

---

🎯 Core Value in One Sentence

PandocPro makes AI practical for serious documents — privately, locally, and without breaking your formatting.

---

🔥 Key Features at a Glance

🧭 Clean Word ↔ Markdown Workflow

PandocPro uses Pandoc under the hood for conversions.
Expect:

• Full structure retention
• Proper headings & lists
• Reliable round-trip exports
• Zero formatting surprises

✂️ Section-based Editing for Long Documents

Instead of feeding a 20,000-token document to an AI:

• Select a section
• Ask for improvements
• Update just that part

That means:

• No token issues
• No broken formatting
• 70–90% fewer tokens per edit
• Higher control, lower risk

🤖 Targeted AI Improvements

Use AI for:

• Tone adjustments
• Grammar, clarity, style polishing
• Summaries or rewrites
• Structural edits or expansions

Each AI call affects only what you select. The rest of the document stays intact.

🛡️ Privacy-first: Local or Cloud LLMs, Your Choice

Work entirely offline using local LLMs (Ollama, LM Studio, Qwen, etc.).
Cloud usage remains optional — only when you choose.

Ideal for:

• Client deliverables under NDA
• Legal, financial, academic, or sensitive documents
• Internal reports or company-only content

🖥️ GUI + CLI + MCP — Work How You Want

GUI

• Intuitive dashboard
• Drag & drop import/export
• Live status of DOCX ⇆ MD sync
• Presets (Academic, Business, Blog)

CLI (for power users)

• dsync tools for automation
• Scriptable workflows

MCP Support

Integrate with MCP-enabled tools (ChatGPT, Claude, etc.) — trigger conversions, edits or exports directly from your AI assistant.

📦 Ready-to-Use Installer for macOS

We provide a .dmg installer — no build steps, no dependency pain.
Download → install → start working.

---

🧑‍💻 Who PandocPro Is For

• Agencies producing proposals, decks, client reports
• Technical writing and documentation teams
• Researchers and academics
• Legal, financial, and consultancy professionals
• Anyone writing or editing 10–100+ page documents
• Anyone wanting AI editing without risking privacy

---

🧩 Feature Overview

Category	Features
Conversion	DOCX ↔ Markdown via Pandoc, template support, clean exports
Editing	Section-based edits, targeted improvements, tone/style changes, summaries
AI	Local & optional cloud LLM support, precise control, safe editing
Workflow	Watch mode, smart sync, conflict handling, export management
Interfaces	GUI, CLI, MCP — flexible for any user style
Privacy	Local-first, optional cloud, no forced uploads
Distribution	Ready-to-use macOS .dmg, scriptable CLI install/build

---

🖼️ Visual Preview

---

🧩 Requirements

PandocPro runs locally on macOS and relies on the following tools:

• Pandoc (required for DOCX ↔ Markdown conversion)

  brew install pandoc
  

• Node.js (required for the CLI, MCP tools, and running the Electron app)

  brew install node
  

• Ollama (optional, but recommended if you want fully private, local AI editing)

  brew install --cask ollama
  ollama pull llama3
  

• Xcode Command Line Tools (optional — only needed if you plan to build PandocPro from source or compile native dependencies)

  xcode-select --install
  

PandocPro works entirely on your device with local LLMs, cloud models, and MCP-enabled editors. Install the required tools above before running the app.

---

📥 Download / Install

✅ macOS (recommended)

Download the latest .dmg installer from the Releases tab — install and run in seconds.

💻 From Source (all platforms)

git clone https://github.com/motacola/PandocPro.git  
cd PandocPro  
npm install  
npm run gui:dev   # for development  
# or  
npm run gui:build  # creates production build (dmg / zip)

---

⚙️ CLI Quick Start

Convert DOCX → Markdown:

dsync convert document.docx

Improve a section:

dsync improve document.md --section 3 --prompt "Tighten tone and improve clarity"

Export back to DOCX:

dsync export document.md --output final-version.docx

---

🔄 Watch & Auto Sync

• Live watch (auto-export on save):

  DOCX_FILE=docs/presentation.docx MD_FILE=docs/presentation.md npm run watch
  

  Use --mode=to-pdf|to-html|to-pptx to change outputs; --once runs a single export and exits.
• Auto mode (newest file wins):

  ./scripts/docx-sync.sh docs/report.docx docs/report.md auto
  

• Keep Markdown + Word twins ready: drop .docx files in docs/, run dsync option 1 once, then leave watch mode running while you edit.

🚢 Release Flow (when to cut a new DMG)

1. Ensure Pandoc + PPTX writer are installed (or CI installs them). Run npm test (smoke + UI server tests).
2. For server-only changes, no DMG is needed; for GUI/electron changes, run:

   npm run gui:build   # production build
   npm run gui:package # opens gui/release with the DMG
   

3. Smoke the drag-and-drop UI: npm run ui and convert a small .md → .docx/.pdf.
4. Tag and publish only when both smoke tests and GUI packaging succeed.

🔗 MCP Quick Start (for LLM-integrated workflows)

Works with Desktop Commander, Context7, and any other MCP-aware client.

# Easiest: one-step installer
./scripts/install-mcp.sh

# Optional: if your client is launched outside the repo, set PROJECT_ROOT
# export PROJECT_ROOT="/absolute/path/to/docx-md-sync"

Restart your MCP client so it discovers the docsync tools, then ask: “Convert report.docx to markdown, improve the introduction, and export to Word.”

Common client setups:

• Desktop Commander (Claude Desktop): restart the app after copying the YAML.
• Context7: reload the workspace/client so it rescans ~/mcp/tools.
• Other MCP clients (VS Code extension, custom CLIs): point them at the same ~/mcp/tools/docsync.yaml; set PROJECT_ROOT if they launch outside the repo.

---

❤️ Why Users Love PandocPro

• Saves time on long-form editing
• Makes AI editing reliable and controlled
• Keeps documents private and local
• Eliminates formatting issues
• Enables professional workflows for serious documents
• Combines editing, AI, and export in one tool

---

📣 Contributing & Feedback

Pull requests and contributions welcome! Create templates, presets, or automation scripts — drop them into /templates or /scripts, and submit a PR.

---

📄 License

MIT