JotNow MCP Server

A built-in local server that connects your notes directly to AI agent workflows.

JotNow includes a built-in local MCP (Model Context Protocol) server that exposes your notes directly to AI agent workflows. No cloud dependency, no signup — enable it in Settings and connect your agent. It adheres to the MCP specifications.

Server URL: http://localhost:37003/mcp

Available Tools

ToolDescription
search_notesSearch notes by text query. Returns matching notes with their ID, name, updated date, and a text snippet.
get_noteGet the full details of a note, including ID, name, dates, content type, pinned status, and text contents.
create_noteCreate a new Markdown note. Returns the new note's ID and name.
update_noteUpdate an existing Markdown note's name and/or contents. Only Markdown notes can be updated.

Setup

1. Enable the MCP Server

Open JotNow → Settings and toggle the local MCP server on.

2. Connect Your Agent

Claude Code
claude mcp add --transport http jotnow http://localhost:37003/mcp

Verify with:

claude mcp list
VS Code (Agent Tools / Copilot)

Add the following to your .vscode/mcp.json (or user-level MCP config):

{
  "servers": {
    "jotnow": {
      "type": "http",
      "url": "http://localhost:37003/mcp"
    }
  }
}
Any MCP-Compatible Client

Point your client at http://localhost:37003/mcp. JotNow implements the standard MCP HTTP transport, so any spec-compliant client will work.

Agent Instructions

To ensure your AI agent automatically uses JotNow for plans, add the following to your agent's system prompt or custom instructions file (AGENTS.md, CLAUDE.md, .github/copilot-instructions.md, etc.):

When you create plans use the *JotNow* tool to manage the plan information.

With this in place, the agent will automatically call create_note or update_note whenever it drafts a plan, keeping all planning artifacts stored directly in JotNow.

Usage Examples

Search notes from a prompt

Search my JotNow notes for anything related to the project roadmap.

The agent calls search_notes with your keywords and returns matching notes.

Use notes as persistent memory

Read the note titled ‘API Design Decisions’ and summarize the constraints.

The agent calls get_note to retrieve and reason over the content.

Write agent output back to JotNow

Save a summary of this conversation as a new note called ‘Meeting Summary’.

The agent calls create_note with the generated content.

Prioritize notes in JotNow

Pin the note.

The agent calls update_note to pin the note in context to the top of the list in the Note Manager.

Remarks

  • The MCP server is local only — it binds to localhost and is never exposed to the internet.
  • The server must be running (JotNow open + MCP enabled in Settings) for agent connections to succeed.
  • All the tools follow standard MCP tool schema conventions (2025-11-25).
  • The agent has access to all notes, but can only create and edit Markdown notes.
← Back to Docs