Unbuilt CLI Documentation

@unbuilt/cli is a command-line tool for analyzing the technologies used on websites. It can detect frameworks, libraries, and other technical details by analyzing the structure and content of web applications.

Installation

npm install -g @unbuilt/cli

Prerequisites

For local (default) analysis mode, Playwright is required:

# Install Playwright browsers
npx playwright install chromium

This is necessary because Unbuilt uses a headless browser to visit websites and analyze their technology stack. The remote analysis mode doesn't require Playwright to be installed locally as it runs on the Unbuilt server.

Commands

Unbuilt CLI provides three main commands:

• analyze: Analyze a single website. (the analyze keyword can be omitted - ex. unbuilt mysite.com)
• batch: Analyze multiple websites from a CSV file
• status: Check the status of an ongoing remote analysis

Analyzing a Single Website

Basic Usage

unbuilt https://example.com

This will analyze the website locally using a headless browser and display the detected technologies.

Options

| Option | Description | |--------|-------------| | -r, --remote | Run analysis remotely via unbuilt server (runs locally by default) | | -s, --save | Saves the analysis results to the database, which serves as a cache for future runs. By default, CLI analyses are executed without caching results. Available only for admins (you need to set UNBUILT_API_KEY env variable when running). | | -n, --async | Enable asynchronous execution - returns a job ID immediately instead of waiting for the result. This ID can be used to check status or retrieve results later with the unbuilt status command. Only available when using --remote. | | -r, --refresh | Force a fresh analysis by bypassing the cache. Ignores any previously saved results and rebuilds the cache with new data. | | -t, --timeout <seconds> | Max time to wait for analysis to complete (default: 120) | | -j, --json | Output results in JSON format | | --session | Use a copy of the current user's Chrome profile data for authentication and session preservation. This allows the analysis to be performed with the user's existing login sessions and cookies without modifying the original profile. The temporary profile copy is automatically cleaned up when the browser context is closed. |

Examples

# Run a local analysis
unbuilt https://example.com

# Run a local analysis with user's session data (logged-in state)
unbuilt https://example.com --session

# Run a remote analysis
unbuilt https://example.com --remote

# Run a local analysis and save the results to DB (cache).
unbuilt https://example.com --save

# Run a local analysis with longer timeout
unbuilt https://example.com --timeout 30000

# Get JSON output
unbuilt https://example.com --json

# Run a remote analysis asynchronously. Will return <analysisId>.
unbuilt https://example.com --remote --async

# Use <analysisId> to get results.
unbuilt status <analysisId>

Analyzing Multiple Websites (Batch Mode)

The batch command allows you to analyze multiple websites from a CSV file.

CSV File Format

Create a CSV file with a list of URLs to analyze. The file should have a column header named url (or the first column will be used if there's no header).

Example websites.csv:

url
https://example.com
https://example.org
https://example.net

Basic Usage

unbuilt batch websites.csv

Options

| Option | Description | |--------|-------------| | -j, --json | Output results in JSON format | | -s, --save | Save the analysis results to the database | | -o, --output <file> | Save results to a specified output file | | -t, --timeout <seconds> | Max time to wait for each analysis (default: 120) | | -c, --concurrent <number> | Maximum number of concurrent analyses (default: 1) |

Examples

# Analyze all websites in the CSV file
unbuilt batch websites.csv

# Analyze multiple websites concurrently (4 at a time)
unbuilt batch websites.csv --concurrent 4

# Save the results to a CSV file
unbuilt batch websites.csv --output results

# Save the results to a JSON file
unbuilt batch websites.csv --output results.json --json

# Analyze websites and save to database
unbuilt batch websites.csv --save

# Set a longer timeout for each analysis
unbuilt batch websites.csv --timeout 300

Checking Analysis Status

When using remote analysis in async mode, you can check the status of an ongoing analysis using the status command.

Basic Usage

unbuilt status <analysisId>

Options

| Option | Description | |--------|-------------| | -j, --json | Output results in JSON format |

Example

unbuilt status abc123-def456-ghi789

User Session Workflow

When using the --session option, the CLI implements the following workflow:

1. Creates a temporary copy of the user's Chrome profile data directory
2. Removes any lock files from the temporary copy
3. Launches a browser context using this temporary profile copy
4. Performs the website analysis with all the user's existing cookies, login sessions, and authentication tokens
5. Automatically cleans up the temporary profile directory when the browser context closes

This approach allows for secure and efficient analysis of authenticated websites without:

• Modifying the user's original Chrome profile
• Requiring users to manually log in during analysis
• Storing sensitive credentials in the CLI tool

The session option is particularly useful for analyzing:

• Authenticated web applications
• Sites with login-only content
• Applications that behave differently for logged-in users

Output Format

By default, Unbuilt displays the analysis results in a human-readable format, showing the detected technologies grouped by category.

When using the --json flag, the output will be in JSON format, which is useful for programmatic processing or integration with other tools.

Performance Considerations

• Local analysis is faster for analyzing a single website.
• Remote analysis is more resource-efficient for your local machine.
• Batch analysis with multiple concurrent processes is the fastest way to analyze many websites.
• Increasing the --concurrent parameter will speed up batch analysis, but requires more system resources.

Using with Database Storage

When using the --save option, analysis results are stored in a database for future reference. This allows for:

• Faster retrieval of previously analyzed websites
• Building historical data about technology usage
• Sharing analysis results with team members

Available only for admins (you need to set UNBUILT_API_KEY env variable when running. Contact [email protected] in order to get access).

Examples of Common Workflows

Quick Technology Check

unbuilt https://example.com

Analysis of Authentication-Required Website

unbuilt https://app.example.com --session

In-depth Analysis with JSON Output for Further Processing

unbuilt https://example.com --json > analysis.json

Batch Analysis of Many Sites with Results Saved to CSV

unbuilt batch websites.csv --concurrent 4 --output tech-analysis

Batch Analysis with Database Storage

UNBUILT_API_KEY="xxxxx" unbuilt batch websites.csv --concurrent 4 --save

Troubleshooting

If you encounter issues with the tool, try the following:

• Ensure you have a stable internet connection
• For batch analysis, start with a smaller number of concurrent processes
• Check that the URL format in your CSV file is correct (should include http:// or https://)
• For local (default) mode, ensure Playwright is properly installed (npx playwright install chromium)
• For websites with complex security measures, try using remote analysis mode instead
• When using --session with the analyze command, ensure Chrome is not currently running to avoid profile lock issues
• In case of loading resources error, check redirects following the resource and try final URL.