brightdata-cli

Imported from brightdata/skills (skills/brightdata-cli/SKILL.md) via skills.sh. Install upstream with npx skills add brightdata/skills --skill brightdata-cli. Copyright stays with the author.

Bright Data CLI

The Bright Data CLI (brightdata or bdata) gives you full access to Bright Data's web data platform from the terminal. It handles authentication, proxy zones, anti-bot bypass, CAPTCHA solving, and JavaScript rendering automatically — the user just needs to log in once.

Installation

If the CLI is not installed yet, guide the user:

macOS / Linux:

curl -fsSL https://cli.brightdata.com/install.sh | bash

Windows or manual install (any platform):

npm install -g @brightdata/cli

Without installing (one-off usage):

npx --yes --package @brightdata/cli brightdata <command>

Requires Node.js >= 20. After install, both brightdata and bdata (shorthand) are available.

First-Time Setup

Before anything else, check if the user is authenticated. If they haven't logged in yet, guide them through the one-time setup:

# One-time login — opens the browser for OAuth, then everything is automatic
bdata login

This single command:

1. Opens the browser for secure OAuth authentication
2. Saves the API key locally (never needs to be entered again)
3. Auto-creates required proxy zones (cli_unlocker, cli_browser)
4. Sets default configuration

After login, every subsequent command works without any manual intervention.

For headless/SSH environments where no browser is available:

bdata login --device

For direct API key authentication (non-interactive):

bdata login --api-key <key>

To verify setup is complete, run:

bdata config

Command Reference

Read references/commands.md for the full command reference with all flags, options, and examples for every command.

Read references/pipelines.md for the pipelines workflow: find a scraper, read its input fields, run it.

Quick Command Overview

bdata is the shorthand for brightdata. Both work identically.

How to Use Each Command

Scraping

Scrape any URL with automatic bot bypass, CAPTCHA handling, and JS rendering:

# Default: returns clean markdown
bdata scrape https://example.com

# Get raw HTML
bdata scrape https://example.com -f html

# Get structured JSON
bdata scrape https://example.com -f json

# Take a screenshot
bdata scrape https://example.com -f screenshot -o page.png

# Geo-targeted scrape from the US
bdata scrape https://amazon.com --country us

# Save to file
bdata scrape https://example.com -o page.md

# Async mode for heavy pages
bdata scrape https://example.com --async

Searching

Search engines with structured JSON output (Google returns parsed organic results, ads, People Also Ask, and related searches):

# Google search with formatted table
bdata search "web scraping best practices"

# Get raw JSON for piping
bdata search "typescript tutorials" --json

# Search Bing
bdata search "bright data pricing" --engine bing

# Localized search
bdata search "restaurants berlin" --country de --language de

# News search
bdata search "AI regulation" --type news

# Extract just URLs
bdata search "open source tools" --json | jq -r '.organic[].link'

Pipelines (Structured Data Extraction)

Run any of thousands of pre-built scrapers. Always look the scraper up first; never guess a dataset_id or field names:

# 1. Find a scraper
bdata pipelines search <site or data type>

# 2. Its methods, then one method's input fields and a ready-to-run command
bdata pipelines info <dataset_id>
bdata pipelines info <dataset_id> <method>

# 3. Run it: URLs as arguments, other input fields as --<field> flags
bdata pipelines run <dataset_id> collect_by_url "<url>"
bdata pipelines run <dataset_id> <discover_method> --<field> <value> --limit-per-input 5

# Output as CSV
bdata pipelines run <dataset_id> <method> "<url>" --format csv -o out.csv

Every record is billed: start with one input. Unknown or missing required fields fail before the run starts.

Checking Status

For long runs started with --async:

bdata pipelines status <snapshot_id>
bdata pipelines status <snapshot_id> --wait --timeout 300
bdata pipelines results <snapshot_id>

Budget & Zones

# Quick account balance
bdata budget

# Detailed balance with pending charges
bdata budget balance

# All zones cost/bandwidth
bdata budget zones

# Specific zone costs
bdata budget zone my_zone

# Date range filter
bdata budget zones --from 2024-01-01T00:00:00 --to 2024-02-01T00:00:00

# List all zones
bdata zones

# Zone details
bdata zones info cli_unlocker

Configuration

# View all config
bdata config

# Set defaults
bdata config set default_zone_unlocker my_zone
bdata config set default_format json

Installing AI Agent Skills

# Interactive picker — choose skills and target agents
bdata skill add

# Install a specific skill
bdata skill add scrape

# List available skills
bdata skill list

Output Modes

Every command supports multiple output formats:

When piped (stdout is not a TTY), colors and spinners are automatically disabled.

Chaining Commands

The CLI is pipe-friendly:

# Search → extract first URL → scrape it
bdata search "top open source projects" --json \
  | jq -r '.organic[0].link' \
  | xargs bdata scrape

# Scrape and view with markdown reader
bdata scrape https://docs.github.com | glow -

# Scraper records to CSV
bdata pipelines run <dataset_id> collect_by_url "<url>" --format csv > out.csv

Environment Variables

These override stored configuration:

Troubleshooting

Key Design Principles

• One-time auth: After bdata login, everything is automatic. No tokens to manage, no keys to pass.
• Zones auto-created: Login creates cli_unlocker and cli_browser zones automatically.
• Smart defaults: Markdown output, auto-detected formats from file extensions, colors only in TTY.
• Pipe-friendly: JSON output + jq for automation. Colors/spinners disabled in pipes.
• Async support: Heavy jobs can run in background with --async + pipelines status --wait.
• npm package: @brightdata/cli — install globally or use via npx.