sdjc84/MCPad
3.2
If you are the rightful owner of MCPad and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to dayong@mcphub.com.
MCPad is a minimal, offline-first rich-text notepad with a built-in MCP server, designed to maintain persistent memory across LLM sessions and facilitate privacy-focused AI journaling.
Tools
5
Resources
0
Prompts
0
MCPad — Persistent AI Notepad with Built-in MCP Server
A minimal, offline-first rich-text notepad whose entire content is exposed as an MCP server.
Close the app — your notes stay available to any local LLM (via Ollama + mcphost, Gemini CLI, Claude Desktop, etc.) forever.
Perfect for:
- Persistent memory across LLM sessions
- “Ask my notepad” workflows with zero token waste
- Privacy-focused AI journaling / knowledge base
https://github.com/user-attachments/assets/your-screenshot-or-gif-here
Why mcpad?
| Feature | mcpad | Obsidian + plugins | Plain text files + RAG |
|---|---|---|---|
| Rich text (bold, lists, etc.) | Yes (TipTap) | Yes | No |
| Persistent MCP tools when app closed | Yes (independent server) | No | No |
| Works with any Ollama model | Yes (via mcphost) | Partial | Partial |
| <20 MB downloadable binary | Yes (Tauri) | Electron bloat | N/A |
| Zero external dependencies | Yes | Many plugins | No GUI |
MVP Features (v0.1)
- Create, edit, delete rich-text notes
- Full-text + semantic search (optional embedding with Ollama/nomic-embed-text)
- Built-in MCP server exposing:
list_notes() → List[NoteMetadata]read_note(id: str) → str(HTML or Markdown)write_note(title: str, content: str) → str(returns id)update_note(id: str, content: str)search_notes(query: str) → List[NoteMetadata]
- Auto-start server on boot (optional tray icon)
- Works out-of-the-box with
mcphost+ Ollama
Quick Start
1. Download binary (Windows / macOS / Linux)
→ Coming after first release — star to get notified!
2. Or run from source (dev)
# Backend MCP server (runs independently)
git clone https://github.com/yourusername/mcpad.git
cd mcpad/server
python -m venv venv && source venv/bin/activate
pip install -r requirements.txt
uvicorn main:app --port 8900 --reload