What is Notelog

Notelog is a command-line tool designed for quickly logging notes, thoughts, todos, insights, and accomplishments. It utilizes the Model Context Protocol (MCP) server to enable interaction with AI assistants.

Use cases

Use cases for Notelog include capturing spontaneous ideas, managing to-do lists, tracking project progress, and allowing AI assistants to log notes and summaries directly through voice or text commands.

How to use

To use Notelog, install it by cloning the repository and building it with ‘cargo install --path .’. You can add notes via command-line arguments, from files, or even through standard input. In MCP mode, it can receive commands from AI assistants to create and manage notes.

Key features

Key features of Notelog include the ability to add notes with various methods (command-line, files, stdin), support for Markdown formatting, organization of notes by year and month, and integration with AI assistants through the MCP server.

Where to use

Notelog can be used in personal productivity, project management, and any scenario where quick note-taking is beneficial. It is particularly useful for individuals who prefer command-line tools and want to integrate note-taking with AI.

Notelog

Notelog is a command-line tool that you can use to record notes as you think of them – thoughts, todos, insights, accomplishments, etc. It includes a Model Context Protocol (MCP) server for use by AI assistants as primary mode of interaction.

Installation

If you’re on Linux (x86-64), you can download a pre-built executable from the Releases page.

To install from source:

1. Clone this repository
2. Build and install with cargo install --path .

Usage

Basic Usage

# Opens an editor to capture a note
notelog

# Add a note with content from command line arguments
notelog add "This is a note" +example-tag

# Add a note with a specific title
notelog add --title "This is a note" "This is the content of the note"

# Add a note from a file
notelog add --title "This is a note" --file /path/to/file

# Add a note with a title and content
notelog foo bar baz --title 'Metasyntactic variables'

# Add a note from stdin
echo "Lorem ipsum" | notelog

# Add a note with a specific notes directory
notelog -d ~/Shanties add -t "Wellerman" There once was a ship

# Opens the most recent note in the editor
notelog last

# Prints the most recent note to stdout
notelog last --print

Notes Directory

[!IMPORTANT]
You will need to create the notes directory yourself before using Notelog.

By default, notes are stored in ~/NoteLog. You can specify a different directory using the -d or --notes-dir option, or by setting the NOTELOG_DIR environment variable.

The notes directory is organized as follows:

~/NoteLog/
├── 2025/
│   ├── 01_January/
│   │   ├── 2025-01-01T17-45 First note.md
│   │   └── 2025-01-02T12-34 Second note.md
│   ├── 02_February/
│   │   └── ...
│   └── ...
├── .notes.db
└── ...

Notelog maintains an SQLite database in the specified notes directory for use as a search index. Notes are monitored for changes and synchronized with the database automatically.

Model Context Protocol Server

Notelog can act as a server that receives commands from AI assistants, allowing you to create, (re-)tag or search notes using natural language (see examples below).

JSON Configuration Example

{
  "mcpServers": {
    "notelog": {
      "command": "notelog",
      "args": [
        "mcp",
        "-d",
        "/path/to/your/NoteLog/directory"
      ]
    }
  }
}

How to set up an MCP server

• Claude Desktop tutorial
• Cursor IDE tutorial

Creating Notes

You can create notes by asking the LLM:

• /log Added Model Context Protocol support to Notelog +mcp +done
• Create a note: "Use this text verbatim in the note"
• Summarize the conversation so far as a notelog

The default prompt instructs the LLM to automatically add tags and a title to the notes as appropriate.

Searching for Notes

You can search for notes using fulltext search or by specific tags:

• Find notes containing "project plan" with tag +important
• Search for notes tagged +sqlite and +til from May 2025
• How many notes tagged +todo do I have?

To avoid bloating the context window too much, a maximum of 25 notes with their IDs will be returned. The LLM can then use the IDs to retrieve the note contents or edit its tags on request.

Editing Tags

You can edit the tags of existing notes by asking the LLM:

• Remove the +draft tag from note abc123 and add +todo
• Mark note def456 as done (this should remove the +todo tag and add the +done tag)

The note IDs can be found by searching for notes as detailed above.