Skip to main content
Version: 1.5.0

Getting Started with thresh on macOS

Complete guide to provision your first container environment on macOS (Apple Silicon) with containerd in under 5 minutes

Beta Support

macOS support is currently in beta. We recommend using containerd with nerdctl for best compatibility. Apple Silicon (M1/M2/M3) is supported.

Architecture Overview​

thresh provides isolated development environments using lightweight containers across multiple platforms:

Prerequisites​

Container Runtime (containerd or Docker)​

Check if containerd/nerdctl or Docker is installed:

# For containerd/nerdctl (Recommended)
nerdctl --version

# Or for Docker
docker --version

Expected output:

nerdctl version 1.0.0 or higher
# OR
Docker version 24.0.0 or higher
Container Runtime Installation

containerd with nerdctl (Recommended):

brew install containerd
brew install nerdctl

Docker Desktop:

brew install --cask docker

After installation, start the services:

# For containerd
sudo brew services start containerd

# For Docker Desktop, launch the application

Installation​

# Download latest release (Apple Silicon)
curl -L https://github.com/dealer426/thresh/releases/latest/download/thresh-macos-arm64.tar.gz -o thresh.tar.gz

# Extract
tar -xzf thresh.tar.gz

# Move to system directory
sudo mv thresh /usr/local/bin/

# Make executable
chmod +x /usr/local/bin/thresh

# Verify installation
thresh version
Binary Size

macOS binaries are ~13 MB (uncompressed) to preserve Apple code signing and notarization. This is larger than the Linux/Windows builds (~5 MB) which use UPX compression.

Option 2: Build from Source​

# Clone repository
cd ~/
git clone https://github.com/dealer426/thresh.git
cd thresh/thresh/Thresh

# Build Native AOT binary for Apple Silicon
dotnet publish -c Release -r osx-arm64 --self-contained

# Copy to system directory
sudo cp bin/Release/net10.0/osx-arm64/publish/thresh /usr/local/bin/

# Make executable
sudo chmod +x /usr/local/bin/thresh

# Verify
thresh version

Configuration​

Set up GitHub Copilot CLI (Required for AI features)​

thresh uses the GitHub Copilot CLI for all AI features.

# Install GitHub Copilot CLI via Homebrew
brew install copilot-cli

# Or via npm
npm install -g @github/copilot

# Or via install script
curl -fsSL https://gh.io/copilot-install | bash

# Launch and authenticate
copilot
# Then type: /login

More Info: https://github.com/github/copilot-cli

AI Model Support

thresh supports 20+ AI models through GitHub Copilot SDK:

  • GPT Models: gpt-4o, gpt-4o-mini, gpt-4-turbo, gpt-4, gpt-3.5-turbo
  • Reasoning Models: o1-preview, o1-mini
  • Claude Models: claude-3.5-sonnet, claude-3.5-haiku, claude-3-opus
  • Gemini Models: gemini-1.5-pro, gemini-1.5-flash
  • Open Source: llama-3.1-405b, mistral-large, and more

Set your preferred model:

thresh config set default-model gpt-4o

Verify Configuration​

thresh config status

Your First Environment​

1. List Available Blueprints​

thresh blueprint list

Example output:

Available blueprints:

alpine-minimal    - Minimal Alpine environment
ubuntu-dev        - Ubuntu development environment with common tools
python-dev        - Python development environment
node-dev          - Node.js development environment
...

2. Provision Your First Environment​

Quick start with Alpine (fastest):

thresh up alpine-minimal

Python development environment:

thresh up python-dev

Ubuntu development environment:

thresh up ubuntu-dev

With verbose output to see progress:

thresh up alpine-minimal --verbose
Performance

Alpine-based environments provision in under 30 seconds on Apple Silicon thanks to:

  • Native AOT compilation (~50ms startup)
  • ARM64 native binary
  • Efficient package management

3. List Your Environments​

# List thresh-managed environments
thresh list

# List nerdctl containers (if using nerdctl)
nerdctl ps -a

# List Docker containers (if using Docker)
docker ps -a

4. Access Your Environment​

# For nerdctl
nerdctl exec -it alpine-minimal /bin/sh

# For Docker
docker exec -it alpine-minimal /bin/sh

5. Remove Environment When Done​

thresh destroy alpine-minimal

AI Features​

Generate Custom Blueprint​

# Generate a blueprint from natural language
thresh blueprint generate "Python data science environment with Jupyter, pandas, and matplotlib" --output data-science

Generated blueprints are automatically saved and available in thresh blueprint list.

Interactive AI Chat​

thresh chat

Example session:

Chat> I need a PHP development environment with nginx and MySQL
AI: Here's a blueprint for PHP development...

Chat> Add Redis to that
AI: Updated blueprint with Redis...

Chat> exit
MCP Server Integration

thresh includes a Model Context Protocol (MCP) server for integration with Claude Desktop, VS Code, and other MCP clients. See the MCP Integration guide for details.

Common Tasks​

View System Metrics​

# Show metrics in text format
thresh metrics

# Export as JSON
thresh metrics --format json

List Available Distributions​

thresh distros

View Configuration​

# View specific setting
thresh config get default-model

# View all configuration
thresh config status

Example Workflows​

Workflow 1: Quick Python Dev Environment​

# Provision Python environment
thresh up python-dev

# Access environment
nerdctl exec -it python-dev /bin/bash

# Inside container:
python3 --version
pip3 --version

# Exit container
exit

# Clean up when done
thresh destroy python-dev

Workflow 2: Generate Custom Environment with AI​

# Generate blueprint
thresh blueprint generate "Go development environment with Docker and PostgreSQL" --output go-dev

# Verify it was saved
thresh blueprint list | grep go-dev

# Provision from custom blueprint
thresh up go-dev

# Access
nerdctl exec -it go-dev /bin/bash

Workflow 3: Create Multiple Test Environments​

# Create test environments
thresh up alpine-minimal
thresh up ubuntu-dev
thresh up node-dev

# List all
thresh list

# Work with specific one
nerdctl exec -it alpine-minimal /bin/sh

# Clean up all
thresh destroy alpine-minimal
thresh destroy ubuntu-dev
thresh destroy node-dev

Troubleshooting​

"containerd not running"​

# Start containerd service
sudo brew services start containerd

# Check status
sudo brew services list | grep containerd

# Restart if needed
sudo brew services restart containerd

"nerdctl: command not found"​

# Install nerdctl
brew install nerdctl

# Verify installation
nerdctl --version

GitHub Copilot CLI Issues​

# Check if Copilot CLI is installed
copilot --version

# Re-authenticate if needed
copilot
# Then: /login

# Verify thresh can access AI
thresh config status

"Container runtime not found"​

# Check if nerdctl is installed
nerdctl --version

# Or check if Docker is installed
docker --version

# Install if needed (see Prerequisites section)

Permission Issues​

# nerdctl may require sudo on some configurations
sudo nerdctl ps

# To run without sudo, configure containerd properly
# See: https://github.com/containerd/nerdctl#rootless-mode
Clear Cache to Start Fresh

If you encounter persistent issues:

# Remove cached files
rm -rf ~/.thresh/cache

# Reset configuration
thresh config reset

# Try again
thresh up alpine-minimal

macOS-Specific Notes​

Apple Silicon (ARM64) Support​

  • thresh is compiled natively for ARM64 (Apple Silicon)
  • Container images must support ARM64/aarch64 architecture
  • Most modern base images (Alpine, Ubuntu, Debian) have ARM64 variants

Rosetta 2 Compatibility​

If you need x86_64 containers:

# Enable Rosetta 2
softwareupdate --install-rosetta

However, native ARM64 containers are recommended for best performance.

Quick Reference Commands​

# Environment Management
thresh up <blueprint>           # Provision environment
thresh list                     # List environments  
thresh list --all              # List all (including stopped)
thresh destroy <name>           # Remove environment

# Blueprint Management
thresh blueprint list          # List available blueprints
thresh blueprint generate <prompt>  # Generate blueprint with AI
thresh blueprint delete <name> # Delete generated blueprint

# AI Features
thresh chat                    # Interactive AI chat

# System
thresh metrics                 # Show performance metrics
thresh config status           # Show configuration status
thresh version                 # Show version

Next Steps​

  1. ✅ Complete installation
  2. ✅ Set up GitHub Copilot CLI authentication
  3. ✅ Provision your first environment
  4. 🎯 Try AI blueprint generation
  5. 🎯 Explore MCP server integration
  6. 🎯 Check out CLI Reference for advanced features
  7. 🎯 Share feedback on GitHub Discussions to help improve macOS support!

Happy provisioning! 🚀