Steve - Cloud-Based Personal Assistant

Steve is a cloud-based personal assistant platform built on AWS using CloudFormation. Unlike Emma (which is built around Claude Code), Steve is designed to integrate with mcpomni-connect, allowing it to interface with multiple language models through a unified architecture. Steve functions as a "head intern" personal assistant with a focus on task automation using various MCP integrations.

Features

• CloudFormation Deployment: One-click deployment to AWS
• Multi-Model Support: Leverage different AI models via mcpomni-connect
• MCP Integration: Connect to Google Workspace, YNAB, Airtable, and more
• Background Processing: Run tasks in the background with status tracking
• Custom Routines: Define and execute multi-step workflows
• Persona Configuration: Customize Steve's behavior and communication style

Quick Start

Option 1: CloudFormation Deployment (Recommended)

1. Deploy CloudFormation Stack:

   aws cloudformation create-stack \
     --stack-name steve \
     --template-body file://infrastructure/cf-template.yaml \
     --parameters \
       ParameterKey=InstanceType,ParameterValue=t4g.medium \
       ParameterKey=KeyName,ParameterValue=your-key-name \
       ParameterKey=UserEmail,ParameterValue=your-email@example.com \
     --capabilities CAPABILITY_IAM
   

2. Connect to Instance:

   ssh -i your-key.pem ec2-user@<instance-ip>
   

3. Start Using Steve:

   steve --help
   

Option 2: Manual Installation

1. Install Dependencies:

   npm install -g mcpomni-connect
   

2. Install Steve CLI:

   npm install -g steve-cli
   

3. Start Using Steve:

   steve --help
   

Usage

# Get help
steve --help

# Ask a question
steve "What files are in this directory?"

# Run a routine
steve --routine=morning

# Use a specific persona
steve --persona=focused "Research AWS cost optimization"

# Run a task in the background
steve --background --print "Analyze this project" --out-file ~/report.txt

# Check status of background processes
steve --status

Demo

Follow the DEMO.md instructions to see Steve in action!

Directory Structure

steve/
├── bin/               # Executable scripts
├── docs/              # Documentation
├── infrastructure/    # CloudFormation templates
├── lib/               # Core libraries
├── scripts/           # Utility scripts
├── src/               # Source code
└── test/              # Test files

Documentation

See the docs/ directory for detailed documentation:

• Design Document
• Implementation Plan
• Installation Guide
• Test Plan
• Emma Comparison

MCP Integration

Steve integrates with multiple MCP (Model Context Protocol) servers:

• Google Workspace (Email, Calendar, Drive)
• YNAB (Budgeting)
• Airtable (Structured data)
• URL Shortener
• Plus any additional MCPs supported by mcpomni-connect

Development

Prerequisites

• Node.js 14+
• npm 6+
• AWS CLI (for deployment)

Setup Development Environment

# Clone the repository
git clone https://github.com/mattweg/steve-assistant.git
cd steve-assistant

# Install dependencies
npm install

# Make scripts executable
chmod +x bin/steve

# Run tests
npm test

License

MIT

Acknowledgments

This project is inspired by the Emma personal assistant project, but offers a different approach by leveraging mcpomni-connect for multi-model support and using CloudFormation for standardized deployment.