Jot

Local AI note-taking CLI. Capture notes in plain English, analyze and categorize them with a local AI model (LM Studio or Ollama), search and summarize your knowledge base. All offline, all private.

What it does

$ jot add "discussed switching from transformers to diffusion models with my advisor"
Jotted: a1b2c3d4

$ jot summarize

=== Jot Summary ===
Total: 1 notes
Today: 1 | This week: 1

Top tags:
  #research: 1
  #advisor: 1

1 action item(s) found:
  1. Finish literature review by March 15

Architecture

jot add "text"  →  SQLite (instant)
               →  background worker → local model (LM Studio or Ollama)
               →  categorize, extract tags, link to related notes

jot search "X"  →  SQLite full-text search
jot tags #Y     →  filtered by tag
jot summarize   →  aggregate stats, action items, tag cloud

Storage: SQLite at ~/.jot/notes.db
Config: ~/.jot/config.json
No API keys. No accounts. No cloud.

Backends

Jot supports both LM Studio and Ollama. Configure in ~/.jot/config.json:

{
  "defaultBackend": "lmstudio",
  "backends": {
    "lmstudio": {
      "url": "http://localhost:1234/v1/chat/completions",
      "model": "qwen3.5-9b",
      "enabled": true
    },
    "ollama": {
      "url": "http://localhost:11434/v1/chat/completions",
      "model": "llama3",
      "enabled": false
    }
  },
  "remote": {
    "enabled": false,
    "url": ""
  }
}

Remote support: Set remote.url to use a model running on a different machine.

Commands

Command	Description
jot add "text"	Capture a note (instant save, async analysis)
jot search "query"	Search notes by content
jot tags [#tag]	List notes by tag, or show all tags
jot list [--raw]	List all notes
jot summarize	Summary: counts, top tags, action items
jot analyze	Run analysis on unanalyzed notes
jot config	Show current configuration
jot help	Show help

Setup

# Clone
git clone https://github.com/lobs-ai/jot.git
cd jot

# Install dependencies
npm install

# Build
npm run build

# Optional: link globally
npm link

# Start taking notes
jot add "my first note"

Google setup

If you want Gmail and Google Calendar integration, create a Google OAuth client JSON first:

1. Go to https://console.cloud.google.com/
2. Create or select a project
3. Enable Gmail API and Google Calendar API
4. Open APIs & Services -> OAuth consent screen
5. Configure the consent screen
6. Add yourself as a test user if the app is still in testing
7. Open APIs & Services -> Credentials
8. Click Create Credentials -> OAuth client ID
9. Choose Desktop app
10. Download the client JSON file

Then connect Jot:

jot google login /path/to/oauth-client.json
jot google gmail --enable
jot google calendar --enable

Notes:

• Use an OAuth client JSON, not a service account key
• Desktop app is the correct client type for the CLI
• If the consent screen is in testing, only listed test users can sign in

Requires: Node.js 18+
Backends: LM Studio or Ollama running locally

Why local models?

Local models are ideal for classification, tagging, and extraction tasks:

• Zero per-query cost
• Private (data never leaves your machine)
• Fast enough for background processing
• Async architecture: note saves instantly, analysis happens after

Stack

• Runtime: Node.js + TypeScript
• Storage: SQLite (better-sqlite3)
• Backends: LM Studio / Ollama (OpenAI-compatible API)
• Architecture: async agent processing, same pattern as lobs-core

Related

lobs-core — Multi-agent orchestration system
lobs-memory — Semantic memory server for AI agents