AI4PKM CLI

A command-line interface for automating personal knowledge management workflows using AI agents. The CLI provides scheduled prompt execution, multi-agent orchestration, and seamless integration with Claude, Gemini, and Codex.

Looking for the Obsidian Vault? The starter PKM vault (templates, workflows, prompts) is in a separate repo: jykim/ai4pkm-vault

• Multi-Agent Orchestrator: Config-driven system that monitors vault files and triggers AI agents automatically
• Cron Job Scheduling: Automated execution of recurring knowledge management tasks
• Multiple AI Agents: Standardized interface to call Claude Code, Gemini CLI, and Codex CLI
• Vault Template Management: Bootstrap new vaults from pre-configured templates
• Self-Update: Keep the CLI up to date with ai4pkm update

• Python 3.8 or higher

pip install -e .

After installation, the CLI is available as the ai4pkm command.

# Show current configuration and usage
ai4pkm

# Start the multi-agent orchestrator
ai4pkm -o

# Execute a prompt directly
ai4pkm -p "What are the key topics in my vault this week?"

# Use a specific agent for one prompt
ai4pkm -a gemini -p "Translate this to Korean: Hello world"

# Start cron scheduler for recurring tasks
ai4pkm -c

# Trigger a specific agent manually
ai4pkm trigger EIC

# Install a starter vault
ai4pkm template install ai4pkm ~/my-vault

# Update CLI to latest version
ai4pkm update

The CLI uses configuration files in the _Settings_/ directory:

See docs/ for detailed documentation:

• CLI Tool Reference - Full command reference and examples
• Orchestrator Guide - Multi-agent orchestrator setup
• Workflows - AI-assisted workflow patterns
• Prompts - Prompt system documentation

• ai4pkm-vault - Starter Obsidian vault for AI4PKM
• Read Jin's PKM in AI Era series for tutorials