mcli-framework 7.19.7

Portable workflow framework - transform any script into a versioned, schedulable command. Store in ~/.mcli/workflows/, version with lockfile, run as daemon or cron job.

MCLI - Portable Workflow Framework

Transform any script into a versioned, portable, schedulable workflow command.

MCLI is a modular CLI framework that lets you write scripts once and run them anywhere - as interactive commands, scheduled jobs, or background daemons. Your workflows live in ~/.mcli/workflows/, are versioned via lockfile, and completely decoupled from the engine source code.

🎯 Core Philosophy

Write a script. Store it. Version it. Run it anywhere. Schedule it. Share it.

No coupling to the engine. No vendor lock-in. Just portable workflows that work.

🚀 Visual Workflow Editing

Edit your workflow JSON files like Jupyter notebooks with our VSCode extension!

Features:

• 📝 Cell-based editing (Jupyter-like interface)
• ⚡ Live code execution (Python, Shell, Bash, Zsh, Fish)
• 🎯 Monaco editor with IntelliSense
• 📊 Rich markdown documentation cells
• 💾 Files stay as .json (git-friendly)

Quick Install:

# From VSCode Marketplace (pending publication)
code --install-extension gwicho38.mcli-framework

# Or install from VSIX
code --install-extension vscode-extension/mcli-framework-2.0.0.vsix

Learn More:

• Extension README - Features and usage
• Installation Guide - Detailed setup
• Workflow Notebooks Docs - Complete guide

⚡ Quick Start

Installation

# Install from PyPI
pip install mcli-framework

# Or with UV (recommended)
uv pip install mcli-framework

Drop & Run: Simplest Way to Add Commands

MCLI automatically converts any script into a workflow command:

# 1. Create a script with metadata comments
cat > ~/.mcli/commands/backup.sh <<'EOF'
#!/usr/bin/env bash
# @description: Backup files to S3
# @version: 1.0.0
# @requires: aws-cli

aws s3 sync /data/ s3://my-bucket/backup/
EOF

# 2. Sync scripts to JSON (auto-runs on startup)
mcli commands sync -g

# 3. Run it!
mcli run -g backup
# Note: 'mcli run' is an alias for 'mcli workflows'

Supported Languages: Python, Bash, JavaScript, TypeScript, Ruby, Perl, Lua

Key Features:

• ✅ Auto-detect language from shebang or extension
• ✅ Extract metadata from @-prefixed comments
• ✅ Keep scripts as source of truth (JSON is auto-generated)
• ✅ File watcher for real-time sync (MCLI_WATCH_SCRIPTS=true)

See Script Sync Documentation for details.

Initialize Workflows Directory

# Initialize workflows in current git repository
mcli init

# Or initialize global workflows
mcli init --global

# Initialize with git repository for workflows
mcli init --git

This creates a .mcli/workflows/ directory (local to your repo) or ~/.mcli/workflows/ (global) with:

• README.md with usage instructions
• commands.lock.json for version tracking
• .gitignore for backup files

Create Your First Workflow

Method 1: From a Python Script

# Write your script
cat > my_task.py << 'EOF'
import click

@click.command()
@click.option('--message', default='Hello', help='Message to display')
def app(message):
    """My custom workflow"""
    click.echo(f"{message} from my workflow!")
EOF

# Import as workflow
mcli commands import my_task.py --name my-task

# Run it
mcli run my-task --message "Hi"

Method 2: Interactive Creation

# Create workflow interactively
mcli new my-task

# Edit in your $EDITOR, then run
mcli run my-task

📦 Workflow System Features

1. Create Workflows

Multiple ways to create workflows:

# Import from existing Python script
mcli commands import script.py --name my-workflow

# Create new workflow interactively
mcli new my-workflow --description "Does something useful"

# List all workflows
mcli commands list

2. Edit & Manage Workflows

# Edit workflow in $EDITOR
mcli edit my-workflow

# Show workflow details
mcli commands info my-workflow

# Search workflows
mcli commands search "pdf"

# Remove workflow
mcli rm my-workflow

3. Export & Import (Portability)

Share workflows across machines or with your team:

# Export all workflows to JSON
mcli commands export my-workflows.json

# Import on another machine
mcli commands import my-workflows.json

# Export single workflow to Python script
mcli commands export my-workflow --script --output my_workflow.py

Your workflows are just JSON files in ~/.mcli/workflows/:

$ ls ~/.mcli/workflows/
pdf-processor.json
data-sync.json
git-commit.json
commands.lock.json  # Version lockfile

4. Version Control with Lockfile

MCLI automatically maintains a lockfile for reproducibility:

# Update lockfile with current workflow versions
mcli lock update

# Verify workflows match lockfile
mcli lock verify

Example commands.lock.json:

{
  "version": "1.0",
  "generated_at": "2025-10-17T10:30:00Z",
  "commands": {
    "pdf-processor": {
      "name": "pdf-processor",
      "description": "Intelligent PDF processor",
      "group": "workflow",
      "version": "1.2",
      "updated_at": "2025-10-15T14:30:00Z"
    }
  }
}

Version control your workflows:

# Add lockfile to git
git add ~/.mcli/workflows/commands.lock.json ~/.mcli/workflows/*.json
git commit -m "Update workflows"

# On another machine
git pull
mcli lock verify  # Ensures consistency

5. IPFS Cloud Sync (Immutable & Free)

Share workflows globally using IPFS - zero configuration, immutable storage:

# Push your workflows to IPFS
mcli workflows sync push -g -d "Production workflows v1.0"
# → Returns: QmXyZ123... (immutable CID)

# Anyone can pull your exact workflow state
mcli workflows sync pull QmXyZ123...

# View sync history
mcli workflows sync history

# Verify a CID is accessible
mcli workflows sync verify QmXyZ123...

Features:

• ✅ Zero config: No accounts or API keys needed
• ✅ Immutable: CID guarantees content authenticity
• ✅ Decentralized: No single point of failure
• ✅ Free forever: Community-hosted IPFS gateways
• ✅ Shareable: Anyone can retrieve via CID

Use Cases:

• Share command sets with team members
• Distribute workflows to community
• Create immutable workflow snapshots
• Backup workflows to decentralized storage

Note: The current implementation uses public IPFS gateways which may have rate limits. For production use, consider running your own IPFS node or using a pinning service like Pinata or web3.storage.

Migration Helper:

Migrate your workflows to IPFS in one command:

# Migrate directory structure AND push to IPFS
mcli self migrate --to-ipfs -d "Production migration"
# → Moves commands/ to workflows/ AND pushes to IPFS

# Just push existing workflows to IPFS
mcli workflows sync push -g -d "Production v1.0"

6. Run as Daemon or Scheduled Task

Workflows aren't coupled to the engine - run them however you want:

As a Daemon:

# Start workflow as background daemon
mcli run daemon start my-task-daemon --workflow my-task

# Check daemon status
mcli run daemon status

# Stop daemon
mcli run daemon stop my-task-daemon

As Scheduled Task:

# Schedule workflow to run every hour
mcli run scheduler add \
  --name hourly-sync \
  --schedule "0 * * * *" \
  --workflow my-task

# List scheduled workflows
mcli run scheduler list

# View logs
mcli run scheduler logs hourly-sync

🎨 Real-World Workflow Examples

Example 1: PDF Processor

# Create PDF processing workflow
mcli commands import pdf_tool.py --name pdf
# Use it
mcli run pdf extract ~/Documents/report.pdf
mcli run pdf compress ~/Documents/*.pdf --output compressed/
mcli run pdf split large.pdf --pages 10

Example 2: Data Sync Workflow

# Create sync workflow
cat > sync.py << 'EOF'
import click
import subprocess

@click.group()
def app():
    """Multi-cloud sync workflow"""
    pass

@app.command()
@click.argument('source')
@click.argument('dest')
def backup(source, dest):
    """Backup data to cloud"""
    subprocess.run(['rclone', 'sync', source, dest])
    click.echo(f"Synced {source} to {dest}")

@app.command()
def status():
    """Check sync status"""
    click.echo("Checking sync status...")
EOF

mcli commands import sync.py --name sync
# Run manually
mcli run sync backup ~/data remote:backup

# Or schedule it
mcli run scheduler add \
  --name nightly-backup \
  --schedule "0 2 * * *" \
  --workflow "sync backup ~/data remote:backup"

Example 3: Git Commit Helper

# Already included as built-in workflow
mcli run git-commit

# Or create your own variant
mcli commands export git-commit --script --output my_git_helper.py
# Edit my_git_helper.py to customize
mcli commands import my_git_helper.py --name my-git

🔧 Workflow Structure

Each workflow is a JSON file with this structure:

{
  "name": "my-workflow",
  "group": "workflow",
  "description": "Does something useful",
  "version": "1.0",
  "metadata": {
    "author": "[e-mail]",
    "tags": ["utility", "automation"]
  },
  "code": "import click\n\[e-mail]()\ndef app():\n    click.echo('Hello!')",
  "updated_at": "2025-10-17T10:00:00Z"
}

🚀 Built-in Workflows

MCLI comes with powerful built-in workflows:

mcli run --help
# or use the full command: mcli workflows --help

Note: mcli run is a convenient alias for mcli workflows

Available workflows:

• pdf - Intelligent PDF processing (extract, compress, split, merge)
• clean - Enhanced Mac system cleaner
• emulator - Android/iOS emulator management
• git-commit - AI-powered commit message generation
• scheduler - Cron-like job scheduling
• daemon - Process management and daemonization
• redis - Redis cache management
• videos - Video processing and overlay removal
• sync - Multi-cloud synchronization
• politician-trading - Now available as standalone package: politician-trading-tracker

💡 Why MCLI?

The Problem

You write scripts. They work. Then:

• ❌ Can't remember where you saved them
• ❌ Hard to share with team members
• ❌ No version control or change tracking
• ❌ Coupling to specific runners or frameworks
• ❌ No easy way to schedule or daemonize

The MCLI Solution

• ✅ Centralized Storage: All workflows in ~/.mcli/workflows/
• ✅ Portable: Export/import as JSON, share anywhere
• ✅ Versioned: Lockfile for reproducibility
• ✅ Decoupled: Zero coupling to engine source code
• ✅ Flexible Execution: Run interactively, scheduled, or as daemon
• ✅ Discoverable: Tab completion, search, info commands

📚 Using MCLI as a Library

MCLI isn't just a CLI tool - it's a powerful Python library for building workflow automation systems!

from mcli.lib.custom_commands import get_command_manager

# Create commands programmatically
manager = get_command_manager()
manager.save_command(
    name="backup",
    code="import click\[e-mail]()...",
    description="Automated backup workflow"
)

# Discover and execute commands
from mcli.lib.discovery.command_discovery import ClickCommandDiscovery
commands = ClickCommandDiscovery().discover_all_commands()

📖 Complete Documentation:

• SDK Documentation - Comprehensive API reference and usage guide
• Library Usage Example - Complete working example
• Custom Commands Guide - Workflow management

Features for Library Users:

• ✅ Command creation and discovery APIs
• ✅ Workflow scheduling and automation
• ✅ Configuration and logging utilities
• ✅ Script synchronization system
• ✅ Performance optimization tools
• ✅ Database and caching integrations
• ✅ Internal utilities (file ops, auth, Redis, LSH client, etc.)

📚 Advanced Features

Shell Completion

# Install completion for your shell
mcli self completion install

# Now use tab completion
mcli run <TAB>                # Shows all workflows
mcli run pdf <TAB>            # Shows pdf subcommands

AI Chat Integration

# Chat with AI about your workflows
mcli chat

# Configure AI providers
export OPENAI_API_KEY=your-key
export ANTHROPIC_API_KEY=your-key

Self-Update

# Update MCLI to latest version
mcli self update

# Check version
mcli version

🛠️ Development

For Development or Customization

# Clone repository
git clone https://github.com/gwicho38/mcli.git
cd mcli

# Setup with UV
uv venv
uv pip install -e ".[dev]"

# Run tests
make test

# Build wheel
make wheel

📖 Documentation

• 📚 Documentation Index: Complete Documentation Index - All docs organized by category
• Installation: See Installation Guide
• Workflows: Full workflow documentation (this README)
• Shell Completion: See Shell Completion Guide
• Testing: See Testing Guide
• Contributing: See Contributing Guide
• Release Notes: See Latest Release (7.16.2)
• Code of Conduct: See Code of Conduct
• Changelog: See Changelog

🎯 Common Use Cases

Use Case 1: Daily Automation Scripts

# Create your daily automation
mcli new daily-tasks  # Add your tasks in $EDITOR
mcli run scheduler add --name daily --schedule "0 9 * * *" --workflow daily-tasks

Use Case 2: Team Workflow Sharing

# On your machine
mcli commands export team-workflows.json

# Share file with team
# On teammate's machine
mcli commands import team-workflows.json
mcli lock verify  # Ensure consistency

Use Case 3: CI/CD Integration

# In your CI pipeline
- pip install mcli-framework
- mcli commands import ci-workflows.json
- mcli run build-and-test
- mcli run deploy --env production

📦 Dependencies

Core (Always Installed)

• click: CLI framework
• rich: Beautiful terminal output
• requests: HTTP client
• python-dotenv: Environment management

Optional Features

All features are included by default as of v7.0.0. For specialized needs:

# GPU support (CUDA required)
pip install "mcli-framework[gpu]"

# Development tools
pip install "mcli-framework[dev]"

🤝 Contributing

We welcome contributions! Especially workflow examples.

1. Fork the repository
2. Create feature branch: git checkout -b feature/awesome-workflow
3. Create your workflow
4. Export it: mcli commands export my-workflow.json
5. Submit PR with workflow JSON

📄 License

MIT License - see LICENSE for details.

🙏 Acknowledgments

• Built with Click
• Styled with Rich
• Managed with UV

---

Start transforming your scripts into portable workflows today:

pip install mcli-framework
mcli new my-first-workflow