Vizzly CLI

Visual review platform for UI developers and designers

What is Vizzly?

Vizzly is a visual review platform designed for how modern teams work. Instead of recreating your components in a sandboxed environment, Vizzly captures screenshots directly from your functional tests. This means you test the real thing, not a snapshot.

It's fast because we don't render anything—we process the images you provide from any source. Bring screenshots from web apps, mobile apps, or even design mockups, and use our collaborative dashboard to streamline the review process between developers and designers.

Features

• 📸 Smart Screenshots - Automatic deduplication and intelligent diffing
• 🎨 Any Screenshot - Web, mobile, desktop, design mockups, or any visual content
• 🏃 TDD Mode - Local visual testing with interactive dashboard, settings, and project tools
• 📊 Beautiful Dashboard - Intuitive web interface for reviewing changes
• 👥 Team Collaboration - Built for UI developers and designers to work together
• 🔄 CI/CD Ready - GitHub, GitLab, CircleCI, and more

Quick Start

Requirements: Node.js 20 or newer.

# Install globally
npm install -g @vizzly-testing/cli

# Initialize your project
vizzly init

Authentication

Vizzly supports two authentication methods:

Option 1: User Authentication (Recommended for local development)

# Authenticate with your Vizzly account
vizzly login

# Optional: Configure project-specific token
vizzly project:select

# Run tests
vizzly run "npm test"

Option 2: API Token (Recommended for CI/CD)

# Set via environment variable
export VIZZLY_TOKEN=your-project-token

# Run tests
vizzly run "npm test"

For local development with .env files:

VIZZLY_TOKEN=your-project-token

Then add .env to your .gitignore file. For CI/CD, use your provider's secret management system.

Token Priority:

1. CLI flag (--token)
2. Environment variable (VIZZLY_TOKEN)
3. Project mapping (configured via vizzly project:select)
4. User access token (from vizzly login)

Upload existing screenshots

vizzly upload ./screenshots --build-name "Release v1.2.3"

Integrate with your tests

# Run tests with cloud integration (CI/CD)
vizzly run "npm test"

# Use TDD mode for local development
vizzly tdd start      # Start TDD server
npm test --watch      # Run tests in watch mode

In your test code

import { vizzlyScreenshot } from '@vizzly-testing/cli/client';

// Option 1: Using a Buffer
const screenshot = await page.screenshot();
await vizzlyScreenshot('homepage', screenshot, {
  browser: 'chrome',
  viewport: '1920x1080'
});

// Option 2: Using a file path
await page.screenshot({ path: './screenshots/homepage.png' });
await vizzlyScreenshot('homepage', './screenshots/homepage.png', {
  browser: 'chrome',
  viewport: '1920x1080'
});

Multi-Language Support: Currently available as a JavaScript/Node.js SDK with Python, Ruby, and other language bindings coming soon. The client SDK is lightweight and simply POSTs screenshot data to the CLI for processing.

Commands

Authentication Commands

vizzly login                        # Authenticate with your Vizzly account
vizzly logout                       # Clear stored authentication tokens
vizzly whoami                       # Show current user and authentication status
vizzly project:select               # Configure project-specific token
vizzly project:list                 # Show all configured projects
vizzly project:token                # Display project token for current directory
vizzly project:remove               # Remove project configuration

Login Command

Authenticate using OAuth 2.0 device flow. Opens your browser to authorize the CLI with your Vizzly account.

# Interactive browser-based login
vizzly login

# JSON output for scripting
vizzly login --json

Features:

• Browser auto-opens with pre-filled device code
• Secure OAuth 2.0 device authorization flow
• 30-day token expiry with automatic refresh
• Tokens stored securely in ~/.vizzly/config.json with 0600 permissions

Logout Command

Clear all stored authentication tokens from your machine.

# Clear all tokens
vizzly logout

# JSON output
vizzly logout --json

Revokes tokens on the server and removes them from local storage.

Whoami Command

Display current authentication status, user information, and organizations.

# Show user and authentication info
vizzly whoami

# JSON output for scripting
vizzly whoami --json

Shows:

• Current user email and name
• Organizations you belong to
• Token status and expiry
• Project mappings (if any)

Project Commands

Configure directory-specific project tokens for multi-project workflows.

# Select a project for current directory
vizzly project:select

# List all configured projects
vizzly project:list

# Show token for current directory
vizzly project:token

# Remove project configuration
vizzly project:remove

Use case: Working on multiple Vizzly projects? Configure each project directory with its specific token. The CLI automatically uses the right token based on your current directory.

Upload Screenshots

vizzly upload <directory>           # Upload screenshots from directory
vizzly upload ./screenshots --wait  # Wait for processing
vizzly upload ./screenshots --upload-all  # Upload all without deduplication
vizzly upload ./screenshots --parallel-id "ci-run-123"  # For parallel CI builds

Run Tests with Integration

vizzly run "npm test"               # Run with Vizzly integration
vizzly run "pytest" --port 3002     # Custom port
vizzly run "npm test" --wait        # Wait for build completion
vizzly run "npm test" --allow-no-token  # Run without API token
vizzly run "npm test" --parallel-id "ci-run-123"  # For parallel CI builds

Run Command Options

Server Configuration:

• --port <port> - Port for screenshot server (default: 47392)
• --timeout <ms> - Server timeout in milliseconds (default: 30000)

Build Configuration:

• -b, --build-name <name> - Custom build name
• --branch <branch> - Git branch override
• --commit <sha> - Git commit SHA override
• --message <msg> - Commit message
• --environment <env> - Environment name (default: test)

Processing Options:

• --wait - Wait for build completion and exit with appropriate code
• --threshold <number> - Comparison threshold (0-1, default: 0.01)
• --upload-timeout <ms> - Upload wait timeout in ms
• --upload-all - Upload all screenshots without SHA deduplication

Parallel Execution:

• --parallel-id <id> - Unique identifier for parallel test execution (also via VIZZLY_PARALLEL_ID)

Development & Testing:

• --allow-no-token - Allow running without API token (useful for local development)
• --token <token> - API token override

Dev Command

For local development with visual testing, use the dev command:

# Start interactive dev server (runs in background)
vizzly dev start

# Run your tests in watch mode (same terminal or new one)
npm test -- --watch

# View the dashboard at http://localhost:47392

Interactive Dashboard: The dev dashboard provides real-time feedback:

• Visual Comparisons - See diffs as tests run with multiple view modes
• Baseline Management - Accept/reject changes directly from the UI
• Configuration Editor - Edit settings without touching config files
• Project Management - Login, link directories to cloud projects
• Test Statistics - Real-time pass/fail metrics
• Dark Theme - Easy on the eyes during long sessions

Dev Subcommands:

# Start the dev server (primary workflow)
vizzly dev start [options]

# Run tests once with ephemeral server (generates static report)
vizzly dev run "npm test" [options]

# Check dev server status
vizzly dev status

# Stop a running dev server
vizzly dev stop

Dev Command Options:

• --set-baseline - Accept current screenshots as new baseline
• --baseline-build <id> - Use specific build as baseline (requires API token)
• --threshold <number> - Comparison threshold (0-1, default: 0.1)
• --port <port> - Server port (default: 47392)
• --timeout <ms> - Server timeout (default: 30000)
• --open - Auto-open dashboard in browser (start command only)

Setup and Status Commands

vizzly init                         # Create vizzly.config.js with defaults
vizzly status <build-id>            # Check build progress and results
vizzly status <build-id> --verbose  # Detailed build information
vizzly status <build-id> --json     # Machine-readable output
vizzly finalize <parallel-id>       # Finalize parallel build after all shards complete
vizzly doctor                       # Fast local preflight (no network)
vizzly doctor --api                 # Include API connectivity checks

Init Command

Creates a basic vizzly.config.js configuration file with sensible defaults. No interactive prompts - just generates a clean config you can customize.

vizzly init           # Create config file
vizzly init --force   # Overwrite existing config

Status Command

Check the progress and results of your builds. Shows comprehensive information including:

• Build status and progress
• Screenshot and comparison counts (new, changed, identical)
• Git branch and commit details
• Direct web dashboard link
• Timing and execution information

# Basic status check
vizzly status abc123-def456-build-id

# Detailed information for debugging
vizzly status abc123-def456-build-id --verbose

# JSON output for CI/CD integration
vizzly status abc123-def456-build-id --json

Doctor

• Purpose: Quickly validate your local setup without network calls by default.
• Checks: Node.js version (>= 20), apiUrl format, comparison threshold, effective port (default 47392).
• Optional: Add --api to verify connectivity using your VIZZLY_TOKEN.

Examples:

# Local-only checks
vizzly doctor

# Include API connectivity
VIZZLY_TOKEN=your-token vizzly doctor --api

# JSON output for tooling
vizzly doctor --json

The dev command provides fast local development with immediate visual feedback. See the Dev Mode Guide for complete details on local visual testing.

Note: The vizzly tdd command is deprecated and will be removed in the next major version. Please use vizzly dev instead.

Configuration

Create a vizzly.config.js file with vizzly init or manually:

export default {
  // Server configuration
  server: {
    port: 47392,
    timeout: 30000
  },

  // Comparison configuration
  comparison: {
    threshold: 0.1,
    ignoreAntialiasing: true
  },

  // Upload configuration
  upload: {
    concurrency: 5,
    timeout: 30000
  }
};

Run vizzly init to generate this file automatically with sensible defaults.

Config Reference

For the full configuration schema and CLI options, see docs/api-reference.md.

Framework Examples

Playwright

import { vizzlyScreenshot } from '@vizzly-testing/cli/client';

test('homepage test', async ({ page }) => {
  await page.goto('/');
  const screenshot = await page.screenshot();
  await vizzlyScreenshot('homepage', screenshot, {
    browser: 'chrome',
    viewport: '1920x1080'
  });
});

Cypress

// cypress/support/commands.js
Cypress.Commands.add('vizzlyScreenshot', (name, properties = {}) => {
  cy.screenshot(name, { capture: 'viewport' });
  cy.readFile(`cypress/screenshots/${name}.png`, 'base64').then((imageBase64) => {
    const imageBuffer = Buffer.from(imageBase64, 'base64');
    return vizzlyScreenshot(name, imageBuffer, {
      browser: Cypress.browser.name,
      ...properties
    });
  });
});

CI/CD Integration

For CI/CD pipelines, use the --wait flag to wait for visual comparison results and get appropriate exit codes:

GitHub Actions

- name: Visual Tests
  run: npx vizzly run "npm test" --wait
  env:
    VIZZLY_TOKEN: ${{ secrets.VIZZLY_TOKEN }}
    # Optional: Provide correct git information from GitHub context
    VIZZLY_COMMIT_MESSAGE: ${{ github.event.pull_request.head.commit.message || github.event.head_commit.message }}
    VIZZLY_COMMIT_SHA: ${{ github.event.pull_request.head.sha || github.event.head_commit.id }}
    VIZZLY_BRANCH: ${{ github.head_ref || github.ref_name }}

Parallel Builds in CI

For parallel test execution, use --parallel-id to ensure all shards contribute to the same build:

# GitHub Actions with parallel matrix
jobs:
  e2e-tests:
    strategy:
      matrix:
        shard: [1/4, 2/4, 3/4, 4/4]
    steps:
      - name: Run tests with Vizzly
        run: |
          npx vizzly run "npm test -- --shard=${{ matrix.shard }}" \
            --parallel-id="${{ github.run_id }}-${{ github.run_attempt }}"
        env:
          VIZZLY_TOKEN: ${{ secrets.VIZZLY_TOKEN }}

  finalize-e2e:
    needs: e2e-tests
    runs-on: ubuntu-latest
    if: always() && needs.e2e-tests.result == 'success'
    steps:
      - name: Finalize parallel build
        run: |
          npx vizzly finalize "${{ github.run_id }}-${{ github.run_attempt }}"
        env:
          VIZZLY_TOKEN: ${{ secrets.VIZZLY_TOKEN }}

GitLab CI

visual-tests:
  stage: test
  image: node:20
  script:
    - npm ci
    - npx vizzly run "npm test" --wait
  variables:
    VIZZLY_TOKEN: $VIZZLY_TOKEN

The --wait flag ensures the process:

• Waits for all screenshots to be processed
• Exits with code 1 if visual differences are detected
• Exits with code 0 if all comparisons pass
• Allows your CI to fail appropriately when visual regressions occur

API Reference

vizzlyScreenshot(name, imageBuffer, properties)

Send a screenshot to Vizzly.

• name (string): Screenshot identifier
• imageBuffer (Buffer | string): Image data as Buffer, or file path to an image
• properties (object): Metadata for organization

File Path Support:

• Accepts both absolute and relative paths
• Automatically reads the file and converts to Buffer internally
• Works with any PNG image file

isVizzlyEnabled()

Check if Vizzly is enabled in the current environment.

AI & Editor Integrations

Claude Code Plugin

Vizzly includes built-in support for Claude Code, Anthropic's official CLI tool. The integration brings AI-powered visual testing workflows directly into your development environment.

Features:

• 🤖 AI-assisted debugging - Get intelligent analysis of visual regressions
• 📊 TDD status insights - Check dashboard status with contextual suggestions
• 🔍 Smart diff analysis - AI helps determine if changes should be accepted or fixed
• ✨ Test coverage suggestions - Get framework-specific screenshot recommendations
• 🛠️ Interactive setup - Guided configuration and CI/CD integration help

Getting Started with Claude Code:

1. Install Claude Code (if you haven't already):

   npm install -g @anthropic-ai/claude-code

2. Install the Vizzly plugin via Claude Code marketplace:

   /plugin marketplace add vizzly-testing/cli
   

3. Use AI-powered workflows with slash commands:

   /vizzly:check-visual-tests      # Check dev server status with AI insights
   /vizzly:debug-visual-regression # Analyze visual failures with AI
   /vizzly:suggest-screenshots     # Find test coverage gaps
   /vizzly:setup                   # Interactive setup wizard
   

The plugin works seamlessly with both local dev mode and cloud builds, providing contextual help based on your current workflow.

Plugin Ecosystem

Vizzly supports a powerful plugin system that allows you to extend the CLI with custom commands. Plugins are automatically discovered from @vizzly-testing/* packages or can be explicitly configured.

Official Plugins

• Claude Code Integration (built-in) - AI-powered visual testing workflows for Claude Code
• @vizzly-testing/storybook (coming soon) - Capture screenshots from Storybook builds

Using Plugins

Plugins under the @vizzly-testing/* scope are auto-discovered:

# Install plugin
npm install @vizzly-testing/storybook

# Use immediately - commands are automatically available!
vizzly storybook ./storybook-static

# Plugin commands show in help
vizzly --help

Creating Plugins

You can create your own plugins to add custom commands:

// plugin.js
export default {
  name: 'my-plugin',
  version: '1.0.0',
  register(program, { config, logger, services }) {
    program
      .command('my-command')
      .description('My custom command')
      .action(async () => {
        logger.info('Running my command!');
      });
  }
};

Add to your vizzly.config.js:

export default {
  plugins: ['./plugin.js']
};

See the Plugin Development Guide for complete documentation and examples.

Documentation

• Getting Started
• Authentication Guide
• Upload Command Guide
• Test Integration Guide
• Dev Mode Guide
• Plugin Development
• API Reference
• Doctor Command

AI & Editor Integrations:

• Claude Code Plugin - Built-in support (see AI & Editor Integrations above)

Environment Variables

Authentication

• VIZZLY_TOKEN: API authentication token (project token or access token). Example: export VIZZLY_TOKEN=your-token.
  • For local development: Use vizzly login instead of manually managing tokens
  • For CI/CD: Use project tokens from environment variables

Core Configuration

• VIZZLY_API_URL: Override API base URL. Default: https://app.vizzly.dev.
• VIZZLY_LOG_LEVEL: Logger level. One of debug, info, warn, error. Example: export VIZZLY_LOG_LEVEL=debug.

Parallel Builds

• VIZZLY_PARALLEL_ID: Unique identifier for parallel test execution. Example: export VIZZLY_PARALLEL_ID=ci-run-123.

Git Information Override

For enhanced CI/CD integration, you can override git detection with these environment variables:

• VIZZLY_COMMIT_SHA: Override detected commit SHA. Useful in CI environments.
• VIZZLY_COMMIT_MESSAGE: Override detected commit message. Useful in CI environments.
• VIZZLY_BRANCH: Override detected branch name. Useful in CI environments.
• VIZZLY_PR_NUMBER: Override detected pull request number. Useful for PR-specific builds.

Example for GitHub Actions:

env:
  VIZZLY_COMMIT_MESSAGE: ${{ github.event.pull_request.head.commit.message || github.event.head_commit.message }}
  VIZZLY_COMMIT_SHA: ${{ github.event.pull_request.head.sha || github.event.head_commit.id }}
  VIZZLY_BRANCH: ${{ github.head_ref || github.ref_name }}
  VIZZLY_PR_NUMBER: ${{ github.event.pull_request.number }}

These variables take highest priority over both CLI arguments and automatic git detection.

Contributing

We welcome contributions! Whether you're fixing bugs, adding features, or improving documentation, your help makes Vizzly better for everyone.

Getting Started

1. Fork the repository on GitHub
2. Clone your fork locally: git clone https://github.com/your-username/cli.git
3. Install dependencies: npm install
4. Run tests to ensure everything works: npm test

Development Workflow

1. Create a feature branch: git checkout -b feature/your-feature-name
2. Make your changes and add tests for any new functionality
3. Run the linter: npm run lint
4. Run tests: npm test
5. Commit your changes using gitmoji format: git commit -m '✨ Add your feature'
6. Push to your fork: git push origin feature/your-feature-name
7. Open a Pull Request

Reporting Issues

Found a bug or have a feature request? Please open an issue with:

• A clear description of the problem or request
• Steps to reproduce (for bugs)
• Your environment details (OS, Node.js version, etc.)

Development Setup

The CLI is built with modern JavaScript and requires Node.js 20+ (LTS). See the development scripts in package.json for available commands.

License

MIT © Stubborn Mule Software