← Back to Supertag CLI

Tana CLI User Guide

Welcome to Tana CLI! This guide will help you get started with the command-line tools for Tana.

Table of Contents

  1. Requirements
  2. Installation
  3. License Activation
  4. Connecting to Tana
  5. First-Time Export Setup
  6. Basic Usage
  7. Advanced Features
  8. Troubleshooting
  9. Quick Reference

Supertag CLI User Guide

Welcome to Supertag CLI! This guide will help you get started with the command-line tools for Tana.

Requirements

Before installing Supertag CLI, make sure you have:

No additional runtime or dependencies are required - Supertag CLI is distributed as standalone executables.

Installation

Step 1: Download Supertag CLI

Download the latest release from GitHub:

Download from GitHub Releases

Or build from source:

git clone https://github.com/jcfischer/supertag-cli.git
cd supertag-cli
bun install
bun run build

Step 2: Extract the Archive (if downloaded)

# Extract the downloaded archive
unzip supertag-cli-*.zip
cd supertag-cli

Step 3: Make Globally Available (Optional)

To run supertag from anywhere:

# Option A: Symlink to /usr/local/bin (recommended)
sudo ln -s $(pwd)/supertag /usr/local/bin/supertag
sudo ln -s $(pwd)/export/supertag-export /usr/local/bin/supertag-export

# Option B: Add to PATH in your shell config
echo 'export PATH="$PATH:/path/to/supertag-cli"' >> ~/.zshrc
source ~/.zshrc

Step 4: Install Playwright for Export Tool

The export tool requires Playwright for browser automation. Install it once:

cd export
npm install

This installs the Playwright browser automation library needed for supertag-export.

Connecting to Tana

Supertag CLI needs your API token to create nodes in Tana.

Step 1: Get Your API Token

  1. Open Tana in your browser
  2. Go to Settings → API (or visit: https://app.tana.inc/?bundle=settings&panel=api)
  3. Click Generate Token
  4. Copy the token

Step 2: Configure the Token

Choose one method:

# Method A: Environment variable (recommended for scripts)
export TANA_API_TOKEN="your-token-here"

# Method B: Save to config file
supertag config --token your-token-here

# Method C: Pass with each command
supertag create todo "Task" --token your-token-here

Verify Connection

supertag config --show

First-Time Export Setup

To search and query your Tana data locally, you need to export and index your workspace.

Step 1: Login and Discover Workspaces

The easiest way to configure workspaces is to let the export tool discover them automatically:

# Login to Tana (opens browser for authentication)
supertag-export login

# Discover all your workspaces
supertag-export discover

# Auto-add discovered workspaces with --add flag
supertag-export discover --add

This automatically captures the rootFileId needed for exports and configures your workspaces.

Step 2: Configure Your Workspace (Manual Alternative)

If you prefer manual configuration:

# Add your workspace with the rootFileId from discover
supertag workspace add M9rkJkwuED --alias main

# Or if you have multiple workspaces
supertag workspace add M9rkJkwuED --alias personal
supertag workspace add 7e25I56wgQ --alias work

# Set your default workspace
supertag workspace set-default main

The login opens a browser window. Log in to Tana as usual (via Google or email). Once you see your workspace, close the browser. Your session is saved for future exports.

Note: On first run, Chromium browser (~300 MB) downloads automatically.

Step 3: Run Your First Export

# Export default workspace
supertag-export run

# Export with verbose output (see auth method used)
supertag-export run -v

Verbose output shows the authentication flow:

[supertag-export] Using cached token
[supertag-export] Token valid (expires in 57 minutes, method: cached)
[supertag-export] Workspace: 🏠 Personal Tana
[supertag-export] Nodes: 47,832, Size: 35.2MB
[supertag-export] Downloading snapshot...
[supertag-export] Saved to: ~/Documents/Tana-Export/personal/M9rkJkwuED@2025-12-09.json

Export is fast - uses Tana's API directly (no browser automation needed):

Export locations by platform:

Step 4: Index the Export

supertag sync index

This creates a local SQLite database for fast searching:

Indexing workspace export...
✓ Indexed 47,832 nodes in 0.4s

Step 5: Verify Setup

# Check database stats
supertag stats

# Try a search
supertag search "meeting"

Basic Usage

Creating Nodes

Create nodes with any supertag directly from the command line:

# Create a todo
supertag create todo "Buy groceries" --status active

# Create a todo with due date
supertag create todo "Submit report" --status active --duedate 2025-12-31

# Create a meeting note
supertag create meeting "Team Standup" --date 2025-12-08

# Create with multiple supertags
supertag create video,towatch "Tutorial Video" --url https://youtube.com/watch?v=xxx

Searching Your Workspace

# Full-text search
supertag search "project planning"

# Limit results
supertag search "AI" --limit 5

# Find by supertag
supertag search --type tagged meeting --limit 10

# Semantic search (requires embeddings)
supertag search --type semantic "productivity tips"

# Show node with depth traversal (children)
supertag nodes show <node-id> -d 3          # 3 levels deep
supertag nodes show <node-id> -d 2 --json   # JSON output with depth

Viewing Statistics

# Database statistics
supertag stats

# Embedding statistics
supertag stats --type embed

# Most used supertags
supertag tags --top --limit 20

# List all supertags
supertag tags list

Advanced Features

Multiple Workspaces

Manage personal and work workspaces separately.

Automatic Discovery (Recommended):

# Login first
supertag-export login

# Discover all workspaces (captures rootFileId automatically)
supertag-export discover

# Auto-add all discovered workspaces
supertag-export discover --add

# Or update existing workspaces with rootFileIds
supertag-export discover --update

Manual Setup:

If you prefer manual configuration, use supertag-export discover first to find your workspace's rootFileId:

# Add workspace with rootFileId (primary identifier for exports)
supertag workspace add M9rkJkwuED --alias personal

# List configured workspaces
supertag workspace list

# Set default workspace
supertag workspace set-default personal

# Query specific workspace
supertag search "budget" -w work

# Export all workspaces
supertag-export run --all

# Export a specific workspace
supertag-export run -w work

Important: Each workspace needs a rootFileId for exports to work. Use supertag-export discover to find it automatically.

Webhook Server

Run a local server for integrations:

# Start server
supertag server start --port 3100

# Start as background daemon
supertag server start --daemon

# Check status
supertag server status

# Stop server
supertag server stop

API Endpoints:

See WEBHOOK-SERVER.md for complete API documentation.

Semantic Search (Vector Embeddings)

Semantic search finds nodes by meaning, not just keywords. Unlike full-text search which requires exact word matches, semantic search understands concepts and context.

Prerequisites:

Semantic search requires the sqlite-vec extension for vector operations. Install it for your platform:

macOS (Homebrew):

brew install asg017/sqlite-vec/sqlite-vec

This installs the extension to /opt/homebrew/lib/sqlite-vec/vec0.dylib (Apple Silicon) or /usr/local/lib/sqlite-vec/vec0.dylib (Intel).

macOS (Manual):

# Download the latest release
curl -L https://github.com/asg017/sqlite-vec/releases/latest/download/sqlite-vec-0.1.6-loadable-macos-aarch64.tar.gz -o sqlite-vec.tar.gz

# Extract
tar xzf sqlite-vec.tar.gz

# Move to library location
sudo mkdir -p /opt/homebrew/lib/sqlite-vec
sudo mv vec0.dylib /opt/homebrew/lib/sqlite-vec/

Linux:

# Download for your architecture
curl -L https://github.com/asg017/sqlite-vec/releases/latest/download/sqlite-vec-0.1.6-loadable-linux-x86_64.tar.gz -o sqlite-vec.tar.gz

# Extract and install
tar xzf sqlite-vec.tar.gz
sudo mkdir -p /usr/local/lib/sqlite-vec
sudo mv vec0.so /usr/local/lib/sqlite-vec/

Windows:

# Download from GitHub releases
# https://github.com/asg017/sqlite-vec/releases

# Extract vec0.dll to a known location, e.g.:
# C:\sqlite-vec\vec0.dll

Supertag CLI automatically searches for the extension in standard locations. To specify a custom path:

# Set custom extension path
export SQLITE_VEC_PATH="/path/to/vec0.dylib"

Setup:

# 1. Install Ollama (recommended provider)
# Download from https://ollama.ai

# 2. Pull the embedding model
ollama pull mxbai-embed-large

# 3. Configure Supertag CLI
supertag embed config --provider ollama --model mxbai-embed-large

# 4. Generate embeddings (one-time, ~5-10 minutes for 300K nodes)
supertag embed generate

# 5. Check status
supertag stats --type embed

Search:

# Basic semantic search
supertag search --type semantic "project planning discussions"

# With more results
supertag search --type semantic "authentication issues" --limit 20

# Include ancestor context
supertag search --type semantic "meeting notes" --include-ancestor

# Minimum similarity threshold
supertag search --type semantic "project ideas" --min-similarity 0.7

# JSON output
supertag search --type semantic "tasks" --format json

How it works:

  1. Embeddings: Each node is converted to a 1024-dimensional vector representing its semantic meaning
  2. Smart Filtering: System automatically excludes noise (timestamps, system nodes, very short text)
    • Short meaningful content like names and concepts ARE included (minimum 3 characters)
    • Entity nodes (tagged items) always embedded regardless of length
  3. Vector Search: Your query is embedded and compared to all node embeddings using cosine similarity
  4. Result Deduplication: Duplicate nodes consolidated (highest similarity kept), but nodes with different tags preserved
  5. Trash Filtering: Deleted nodes with _TRASH ancestors are filtered out

CLI Results Format:

Semantic search returns Tana Paste references using [[Name^nodeID]] syntax. These are proper references to existing nodes - clicking them in Tana will navigate to the actual node (not create duplicates).

Webhook API Results:

The webhook server returns results as a formatted Tana table with three columns:

- Semantic Search Results %%view:table%%
  - [[Found Node Name^nodeID]]
    - Ancestor:: [[Parent Node^parentID]]
    - Similarity:: 85%
  - [[Another Result^nodeID2]]
    - Ancestor:: [[Parent^parentID2]]
    - Similarity:: 72%

Each result includes:

Webhook API Examples:

# Semantic search via HTTP (returns Tana table format)
curl -X POST http://localhost:3100/semantic-search \
  -H "Content-Type: application/json" \
  -d '{"query": "productivity tips", "limit": 10}'

# JSON format response (for programmatic access)
curl -X POST http://localhost:3100/semantic-search \
  -H "Content-Type: application/json" \
  -d '{"query": "productivity tips", "limit": 10, "format": "json"}'

Known Limitation:

Some deleted nodes may occasionally appear in results because Tana's JSON export doesn't include comprehensive deletion metadata. The tool filters nodes with _TRASH ownership patterns, but cannot detect all deletion states from exports alone.

Schema Registry

View and use your custom supertags:

# Sync schema from latest export
supertag schema sync

# List all supertags
supertag schema list

# Show fields for a supertag
supertag schema show todo

# Search supertags
supertag schema search project

Export Cleanup

Remove old export files to save disk space:

# Show what would be deleted (dry run)
supertag sync cleanup --dry-run

# Delete old exports, keep last 7 (default)
supertag sync cleanup

# Keep custom number of files
supertag sync cleanup --keep 5

# Clean up all workspaces
supertag sync cleanup --all

Configure in ~/.config/supertag/config.json:

{
  "cleanup": {
    "keepCount": 7,
    "autoCleanup": false
  }
}

Set autoCleanup: true to automatically clean up after each sync.

MCP Server (AI Integration)

The MCP (Model Context Protocol) server enables Tana integration with AI tools like ChatGPT Desktop, Cursor, VS Code Copilot, Claude Code, and other MCP-compatible applications. MCP is the emerging industry standard for AI-tool integration, supported by Anthropic, OpenAI, Google, and Microsoft.

What is MCP?

Think of MCP as "USB-C for AI" - a universal standard that lets any AI assistant access your Tana data. The supertag-mcp binary runs locally on your machine as a subprocess - no server setup, no cloud hosting, no network exposure needed.

Available Tools

Tool Description Example Use
tana_search Full-text search across workspace "Search my Tana for project planning notes"
tana_semantic_search Vector similarity search (requires embeddings) "Find notes about knowledge management"
tana_tagged Find nodes by supertag "Find all my todos"
tana_stats Database statistics "How many nodes are in my Tana?"
tana_supertags List all supertags "What supertags do I have?"
tana_node Get node contents with depth "Show me node abc123 with its children"
tana_create Create new nodes with references "Create a todo linked to node abc123"
tana_sync Trigger reindex or check status "Reindex my Tana database"

Important: Creating References

The Tana Input API does not support inline reference syntax ([[text^nodeId]]) in node names. If you ask an AI to create a node "with a link to X", the AI should use the children parameter, not inline syntax.

Correct way to create a node with a reference:

{
  "supertag": "todo",
  "name": "Follow up on feedback form",
  "children": [{"name": "NetSec Feedback Form", "id": "dvpAO46vtnrx"}]
}

This creates a todo with a child reference that links to the existing node.

Prerequisites

Before setting up MCP, ensure you have:

  1. Indexed database: Run supertag sync index at least once
  2. API token (for tana_create): Configure via supertag config --token YOUR_TOKEN
  3. Schema registry (for tana_create): Run supertag schema sync

Setup for Claude Desktop

Claude Desktop was the first app to support MCP (November 2024).

macOS: Edit ~/Library/Application Support/Claude/claude_desktop_config.json

Windows: Edit %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "tana": {
      "command": "/path/to/supertag-mcp",
      "env": {
        "TANA_WORKSPACE": "personal"
      }
    }
  }
}

After saving, restart Claude Desktop. You should see Tana tools available in the tools menu.

Setup for Claude Code (CLI)

Add to ~/.claude/mcp.json:

{
  "mcpServers": {
    "tana": {
      "command": "/path/to/supertag-mcp",
      "env": {
        "TANA_WORKSPACE": "personal"
      }
    }
  }
}

Restart Claude Code to load the MCP server.

Setup for ChatGPT Desktop

ChatGPT Desktop added MCP support in March 2025.

macOS: Edit ~/Library/Application Support/ChatGPT/chatgpt_config.json

Windows: Edit %APPDATA%\ChatGPT\chatgpt_config.json

{
  "mcpServers": {
    "tana": {
      "command": "/path/to/supertag-mcp"
    }
  }
}

Restart ChatGPT Desktop after saving.

Setup for Cursor IDE

Cursor has excellent MCP support for code-focused workflows.

  1. Create/edit ~/.cursor/mcp.json:
{
  "mcpServers": {
    "tana": {
      "command": "/path/to/supertag-mcp"
    }
  }
}
  1. Open Cursor Settings → Features → MCP
  2. Enable MCP and verify the Tana tools appear
  3. Restart Cursor

Setup for VS Code with Copilot

VS Code supports MCP through the GitHub Copilot extension.

Edit .vscode/mcp.json in your project or user settings:

{
  "mcpServers": {
    "tana": {
      "command": "/path/to/supertag-mcp"
    }
  }
}

Setup for Windsurf

Windsurf (Codeium's AI IDE) supports MCP natively.

Add to Windsurf settings (Settings → MCP Servers):

{
  "tana": {
    "command": "/path/to/supertag-mcp"
  }
}

Windows Configuration

On Windows, use the full path to the executable:

{
  "mcpServers": {
    "tana": {
      "command": "C:\\path\\to\\supertag-mcp.exe"
    }
  }
}

Environment Variables

Variable Description Default
TANA_WORKSPACE Workspace alias to use Default workspace
DEBUG Enable debug logging false
SERVICE_NAME Service name for logs supertag-mcp

Testing Your MCP Setup

Use the MCP Inspector to test your server:

npx @modelcontextprotocol/inspector /path/to/supertag-mcp

This opens a web interface where you can:

Example AI Conversations

Searching:

"Search my Tana for notes about authentication" → Uses tana_search to find relevant nodes

Finding by tag:

"Show me all my meeting notes from this week" → Uses tana_tagged with the meeting supertag

Creating nodes:

"Create a todo to review the pull request" → Uses tana_create to add a new todo node

Exploring structure:

"What supertags do I have in my workspace?" → Uses tana_supertags to list all tags

Database maintenance:

"Reindex my Tana database" → Uses tana_sync to trigger reindexing

How It Works

  1. Local execution: The MCP server runs entirely on your machine
  2. Subprocess model: AI tools spawn it as needed (like running any CLI)
  3. stdio communication: Uses stdin/stdout JSON-RPC (no network ports)
  4. Database access: Reads your local SQLite database
  5. API access: Write operations use your Tana API token
  6. 100% private: All data stays on your machine

Troubleshooting MCP

Tools not appearing:

  1. Check the config file path is correct for your OS
  2. Verify the supertag-mcp path is absolute
  3. Restart the AI application completely
  4. Check logs: tail -f ~/.local/state/supertag/logs/*.log

"Database not found" errors:

supertag sync index  # Index your export first

"Schema registry empty" errors (for tana_create):

supertag schema sync  # Sync schema from export

"API token not configured" errors (for tana_create):

supertag config --token YOUR_TOKEN

Debug mode:

{
  "mcpServers": {
    "tana": {
      "command": "/path/to/supertag-mcp",
      "env": {
        "DEBUG": "true"
      }
    }
  }
}

Setup for Local LLMs (Ollama)

You can use Supertag MCP tools with local LLMs running in Ollama. This provides completely offline, private AI access to your Tana data - no cloud API required.

Why use Ollama?

Prerequisites:

  1. Ollama installed and running
  2. A tool-calling capable model (7B+ parameters recommended)
  3. mcphost - MCP client for Ollama

Step 1: Install mcphost

# macOS (recommended)
brew install mcphost

# Or build from source (requires Go)
go install github.com/mark3labs/mcphost@latest

Step 2: Configure MCP servers

Create ~/.mcp.json (mcphost's config file):

{
  "mcpServers": {
    "supertag": {
      "command": "/path/to/supertag-mcp",
      "args": []
    },
    "datetime": {
      "command": "/path/to/datetime-mcp",
      "args": []
    }
  }
}

Replace the paths with actual paths to your binaries.

Why datetime-mcp? Local LLMs don't have real-time awareness - they don't know the current date or time. The datetime MCP provides get_current_datetime and get_date_info tools so your local LLM can answer questions like "what day is it?" or "show my meetings for today".

Step 3: Pull a tool-calling model

Not all models support tool calling well. Recommended models:

# Recommended - fast and reliable tool calling
ollama pull qwen2.5:7b

# Alternatives
ollama pull llama3.1:8b
ollama pull mistral:7b

Note: Smaller models (1B-3B) may struggle with tool calling. 7B+ parameter models work best.

Step 4: Run mcphost

Interactive mode (chat with your Tana data):

mcphost --quiet -m ollama:qwen2.5:7b

Tip: Use --quiet to reduce visual noise from MCP server initialization messages.

Now you can chat naturally:

> What's the date today?
> What supertags do I have?
> Find all my todos
> Search for notes about authentication

Single prompt mode (for scripts):

mcphost --quiet -m ollama:qwen2.5:7b -p "List the top 5 supertags in workspace 'main'"

Example output:

Executing supertag__tana_supertags...

The top 5 supertags in the 'main' workspace are:
1. meeting (2190 nodes)
2. todo (1847 nodes)
3. contact (956 nodes)
4. project (543 nodes)
5. note (412 nodes)

Supported Models:

Model Tool Calling Speed Notes
qwen2.5:7b Excellent Fast Recommended - best balance
qwen2.5:14b Excellent Medium More capable, needs more RAM
llama3.1:8b Good Fast Good alternative
mistral:7b Fair Fast Works but less reliable
deepseek-r1:7b Good Medium Good reasoning

Model Comparison Test Results

We tested multiple models with the query "List all my meetings in workspace 'main' for today" to evaluate MCP tool-calling capabilities and date awareness.

Test Results (December 2025):

Model Tool Execution Date Awareness Result
Claude Code ✅ Correct ✅ Real-time 5 meetings - Uses actual date (2025-12-11)
qwen2.5:7b ✅ Works ❌ Training cutoff Returns wrong results without explicit date
llama3.1:8b ✅ Works ❌ Training cutoff Returns wrong results without explicit date
qwen2.5:14b ❌ Failed N/A Outputs tool call as text
mistral:7b ❌ Failed N/A Recognizes but doesn't execute

Critical Finding: Date Awareness

The key difference between Claude Code and local LLMs is date awareness:

When told to use datetime tool first:

"First get the current date using the datetime tool, then list my meetings for that date"

qwen2.5:7b correctly chains: get_current_datetime → gets "2025-12-11" → tana_tagged → 5 meetings ✅

Without explicit instruction:

"List meetings for today"

Local LLMs skip the datetime tool and use training cutoff dates → wrong results ❌

Today's 5 Meetings (December 11, 2025):

  1. SAVE THE DATE - Weihnachtsapéro
  2. CoP ACs
  3. MIttagessen Network-/Security Platz Zürich
  4. CYPHR Tension Meeting
  5. Michael Mahlberg's Birthday

Recommendations:

  1. For accurate "today" queries: Use Claude Code or provide explicit dates
  2. For local LLMs: Always include the date in your prompt, or use get_current_datetime first
  3. Best local model: qwen2.5:7b - most reliable tool execution when given proper parameters

Troubleshooting Ollama:

"Model doesn't call tools"

"mcphost not found"

"Connection refused"

Automated Daily Sync

Keep your local database up-to-date:

# Manual: Export and index
supertag-export run && supertag sync index

# Or use the daily script (includes cleanup)
./tana-daily

# Skip cleanup
./tana-daily --no-cleanup

# Cleanup only
./tana-daily --cleanup

macOS LaunchAgent (runs daily at 6 AM):

cp launchd/ch.invisible.tana-daily.plist ~/Library/LaunchAgents/
launchctl load ~/Library/LaunchAgents/ch.invisible.tana-daily.plist

Troubleshooting

"API token not configured"

# Set your token
supertag config --token your-token-here

# Verify
supertag config --show

"Database not found"

You need to export and index first:

supertag-export login    # First time only
supertag-export run      # Export workspace
supertag sync index      # Index the export

"Chromium not found"

Chromium auto-installs on first run. If it fails:

cd export
npx playwright install chromium

Export Hangs or Fails

  1. Check export status: supertag-export status
  2. Try logging in again: supertag-export login
  3. Check your internet connection
  4. Ensure Tana is accessible in your browser

Auth troubleshooting (use -v flag to see auth method):

supertag-export run -v

Output shows which auth method was used:

If exports consistently fail, clear the auth cache and re-login:

rm ~/.cache/supertag/auth-token.json
supertag-export login

Search Returns No Results

  1. Verify database exists: supertag stats
  2. Re-index if export was updated: supertag sync index
  3. Check your search query (use simple terms)

"sqlite-vec extension not found"

The sqlite-vec extension is required for semantic search. Install it:

# macOS (Homebrew - recommended)
brew install asg017/sqlite-vec/sqlite-vec

# Verify installation
ls /opt/homebrew/lib/sqlite-vec/vec0.dylib

If installed in a non-standard location, set the path:

export SQLITE_VEC_PATH="/path/to/vec0.dylib"

See Semantic Search Prerequisites for full installation instructions.

View All Configuration Paths

supertag paths

macOS Output:

📁 Supertag CLI Paths:

   Config:     ~/.config/supertag/
   Data:       ~/.local/share/supertag/
   Cache:      ~/.cache/supertag/
   Database:   ~/.local/share/supertag/supertag-index.db
   Exports:    ~/Documents/Tana-Export/

Linux Output:

📁 Supertag CLI Paths:

   Config:     ~/.config/supertag/
   Data:       ~/.local/share/supertag/
   Cache:      ~/.cache/supertag/
   Database:   ~/.local/share/supertag/supertag-index.db
   Exports:    ~/.local/share/supertag/exports/

Note: Uses supertag namespace to avoid conflicts with the official Tana app.

Note: Supertag CLI follows the XDG Base Directory Specification on all platforms. You can override paths using environment variables like XDG_CONFIG_HOME, XDG_DATA_HOME, etc.

Getting Help

Command Help

supertag --help
supertag search --help
supertag create --help

Support

Contributing

Supertag CLI is open source! Contributions are welcome via pull requests on GitHub.

Quick Reference

Task Command
Set API token supertag config --token TOKEN
Discover workspaces supertag-export discover
Add workspace supertag workspace add ROOTFILEID --alias NAME
Set default workspace supertag workspace set-default NAME
List workspaces supertag workspace list
Login for export supertag-export login
Export workspace supertag-export run
Export (verbose) supertag-export run -v
Export status supertag-export status
Index export supertag sync index
Cleanup old exports supertag sync cleanup
Search
Full-text search supertag search "term"
Semantic search supertag search --type semantic "concept"
Find by tag supertag search --type tagged meeting
Nodes
Show node supertag nodes show <id>
Show with children supertag nodes show <id> -d 3
Tags
List supertags supertag tags list
Top supertags supertag tags --top
Stats
Database stats supertag stats
Embedding stats supertag stats --type embed
Other
Create todo supertag create todo "Task" --status active
Start server supertag server start --daemon
View paths supertag paths
MCP Server
Test MCP server npx @modelcontextprotocol/inspector supertag-mcp

Supertag CLI is open source software, developed by InVisible GmbH, Switzerland. View on GitHub

Tana CLI is developed by InVisible GmbH, Switzerland.