Middleman CLI

Command-line interface for the Middleman GPU compute platform.

Installation

pip install middleman-cli

Quick Start

# Authenticate with your API key
middleman login

# Submit a training job
middleman run train.py --gpu a100

# Check job status
middleman jobs list
middleman jobs status <job-id>

# View logs
middleman jobs logs <job-id> --follow

# Check credit balance
middleman credits

Commands

Authentication

# Login with API key
middleman login

# Show current user
middleman whoami

# Logout
middleman logout

Running Jobs

# Basic job submission
middleman run train.py --gpu t4

# Advanced options
middleman run train.py \
  --gpu a100 \
  --name "my-experiment" \
  --framework pytorch \
  --max-runtime 8 \
  --env WANDB_API_KEY=xxx \
  --requirements requirements.txt \
  --wait

# GPU options: t4, v100, a100, a100-80gb, h100
# Framework options: pytorch, tensorflow, jax, custom

Managing Jobs

# List jobs
middleman jobs list
middleman jobs list --status running
middleman jobs list --all

# Get job status
middleman jobs status <job-id>
middleman jobs status <job-id> --watch

# View logs
middleman jobs logs <job-id>
middleman jobs logs <job-id> --tail 200
middleman jobs logs <job-id> --follow

# Cancel job
middleman jobs cancel <job-id>

Credits

# Show credit balance
middleman credits

Configuration

# Show config
middleman configure --show

# Set custom API URL
middleman configure --api-url https://custom.api.url

# Reset config
middleman configure --reset

Configuration

Configuration is stored in ~/.middleman/config.yaml:

api_key: mdlm_xxxxx
api_url: https://api.middleman.run/api/v1

Environment Variables

• MIDDLEMAN_API_KEY - API key (overrides config file)
• MIDDLEMAN_API_URL - API URL (overrides config file)

Get Your API Key

1. Sign up at https://middleman.run
2. Go to https://app.middleman.run/dashboard/api-keys
3. Create a new API key
4. Run middleman login

Support

• Documentation: https://docs.middleman.run
• Email: support@middleman.run
• Dashboard: https://app.middleman.run