Skip to Content
CLICLI Quick Start

CLI Quick Start

Get started with OptixLog CLI in minutes.

First Command

Check that the CLI is installed:

optixlog --help

You should see:

Usage: optixlog [options] [command]

OptixLog CLI - Experiment tracking

Options:
  -V, --version   output the version number
  -h, --help      display help for command

Commands:
  init           Initialize OptixLog configuration
  config         Manage OptixLog configuration
  add-logging    Auto-instrument Python code with OptixLog
  runs           List recent runs
  check-sdk      Check if OptixLog SDK is installed
  install-sdk    Install OptixLog SDK
  help [command] display help for command

Initialize Configuration

Set up your OptixLog configuration:

optixlog init --project "MyProject" --api-key "proj_YOUR_KEY"

This creates a .optixlog.json file in your current directory (or home directory) with:

{
  "api_key": "proj_YOUR_KEY",
  "api_url": "https://backend.optixlog.com",
  "project": "MyProject"
}

Get your API key: https://optixlog.com 

Your First Transformation

Create a simple Python script:

# script.py
import numpy as np
import matplotlib.pyplot as plt
 
x = np.linspace(0, 10, 100)
y = np.sin(x)
 
plt.plot(x, y)
plt.title("Sine Wave")
plt.savefig("plot.png")

Add OptixLog logging:

optixlog add-logging script.py

The CLI will:

  1. Create a backup: script_backup.py
  2. Transform the code to include OptixLog
  3. Show you what was added

Transformed code:

import optixlog
import os
import numpy as np
import matplotlib.pyplot as plt
 
# Initialize OptixLog
with optixlog.run(
    run_name="experiment",
    api_key=os.getenv("OPTIX_API_KEY"),
    project=os.getenv("OPTIX_PROJECT", "MyProject"),
    config={},
    create_project_if_not_exists=True
) as client:
    x = np.linspace(0, 10, 100)
    y = np.sin(x)
    
    plt.plot(x, y)
    plt.title("Sine Wave")
    
    # Log the plot
    fig = plt.gcf()
    client.log_matplotlib("plot", fig)
    
    plt.savefig("plot.png")

Run Your Code

Execute the instrumented script:

python script.py

The script will:

  1. Initialize OptixLog
  2. Run your code
  3. Log metrics and plots automatically
  4. Show colored output in terminal

View Results

List your runs:

optixlog runs

Output:

🔍 Fetching runs...

Found 1 runs:

✓ experiment
  ID: run_abc123
  Project: MyProject
  Created: 2025-11-24T10:30:00Z

View results in the web dashboard: https://optixlog.com 

Smart Mode

Use --smart flag for better code generation:

optixlog add-logging script.py --smart

Smart mode:

  • Uses SDK helper methods (log_matplotlib())
  • Wraps code in context manager
  • Adds helpful comments
  • Works locally (no backend needed)

Next Steps

CLI InstallationCLI Troubleshooting