aNotepad

Name: aNotepad
Author: anotepad

@anotepadMCP server · External

AI access to your aNotepad online notes: read, search, write, and organize via 22 tools.

Note takingv1.0.0MIT

Freeno charge per call

One address for any MCP client. Your key stays in your Odel vault.

What it does

Reads
Writes
High impact

At most: moves money, runs code or commands, deletes data, or sends messages for you.

What it costs

Free

Odel charges nothing per call. The publisher may charge for its own service.

What it asks for at sign-in

Access to	Read	Change
Notepad	Not stated: notepad.notes, notepad.collections, notepad.comments, notepad.files

From the sign-in at anotepad.com, as the endpoint publishes it. You approve it there.

Tools

None listed · checked 9 Oct 2026

Odel could not list its tools without signing in.

Try it in the Inspector

README3 min readView on GitHub

aNotepad MCP Server

Connect your aNotepad online notes to Claude, ChatGPT, and Cursor through the Model Context Protocol. Read, search, write, and organize your notes through natural conversation.

Documentation · anotepad.com · Support

Note: This is the public home for the aNotepad MCP server. aNotepad is a remote, hosted MCP server — there's nothing to install or run. You connect your AI client straight to the hosted endpoint below and sign in with your aNotepad account. The server source lives in the private aNotepad codebase; this repo holds the registry manifest, docs, and branding.

Connection


Endpoint	`https://api.anotepad.com/mcp`
Transport	Streamable HTTP
Auth	OAuth 2.1 (PKCE) + Dynamic Client Registration (RFC 7591) — no API key setup
Required plan	aNotepad Pro or Max

Quick start

Make sure your aNotepad account is on the Pro or Max plan.
In your AI client's settings, add an MCP server with URL https://api.anotepad.com/mcp.
The client opens a browser window to sign in to aNotepad and approve access.
Ask things like "summarize my recipe collection" or "find the note where I saved that GitHub link."

Client setup

Claude Desktop

Settings → Connectors (or Settings → Developer → Edit Config on older versions).
Add custom connector, paste https://api.anotepad.com/mcp.
Sign in to aNotepad in the browser window and approve.
The aNotepad tools appear under the connectors list.

ChatGPT (Plus / Pro / Team / Enterprise)

Your name (bottom left) → Settings → Connectors → Connect more.
Add custom connector, paste https://api.anotepad.com/mcp.
Approve the OAuth flow.

Cursor

Settings → MCP → Add new MCP server.
Transport HTTP / SSE, URL https://api.anotepad.com/mcp.
Save, then sign in and approve in the browser window.

MCP Inspector (testing)

bash

npx @modelcontextprotocol/inspector

Set transport to Streamable HTTP, URL https://api.anotepad.com/mcp, then Connect.

Tools

22 tools across six groups.

Group	Tools
Notes	`list_notes`, `get_note`, `search_notes`, `create_note`, `update_note`, `delete_note`
Versions	`list_note_versions`, `get_note_version`, `restore_note_version`
Export & email	`get_note_download_url`, `email_note`
Collections	`list_collections`, `create_collection`, `move_note_to_collection`, `delete_collection`
Comments	`list_comments`, `list_recent_comments`, `create_comment`, `delete_comment`
Files & bookmarks	`list_note_files`, `get_note_file_download_url`, `delete_note_file`, `list_bookmarks_in_note`, `add_bookmark`, `remove_bookmark`

What's not exposed

Some operations stay on the web app, by design:

Setting note passwords — removing a password is fine; setting a new value via AI is blocked (it would land in chat transcripts).
File uploads — AI can list, download, and delete attachments, but new uploads go through the web app.
Account / billing changes — plan changes, payment methods, etc. are web-app only.
Creator wallet setup — switching a note to Paid is supported, but the one-time wallet/email verification must be done in the web app first.

Privacy & security

The AI sees only your own notes and their comments — never other users' notes or your account/billing data.
Comments on public notes are wrapped in <untrusted-comment> markers in tool responses, signaling to the AI that the text is data, not instructions — a defense against prompt injection.
Destructive tools (delete, change access, send email) are marked destructive, so most clients require an extra approval prompt.
Note content and passwords are never written to server logs.