Bright Data Web MCP

Official

Real-time, unblocked public web access for AI

How to install ↓ Compare GitHub source repository ↗

Category: Other #88 of 230
Stars: ★ 2.7k Very popular
Transport: stdio (local process)
Runtime: Node.js · Hosted (no local runtime)
Credentials: API key / credential required
License: MIT
Last commit: 9 days ago
Tools: 29

58FMRS · C

Bright Data Web MCP is an MIT-licensed, stdio-based public web access server offering search, scraping, browser automation, and structured data tools. It suits real-time web research and AI-agent workflows, but requires a Bright Data API_TOKEN and involves credit consumption and possible Pro usage charges.

Strongest · Documentation 15/20 Weakest · Security and permissions 8/20

Reliability

9/20

Security and permissions

8/20

Maintenance

12/20

Documentation

15/20

Setup experience

14/20

Why each score

Reliability 9/20: The manifest declares an npm stdio package, version, and required API_TOKEN, and the README provides local and hosted connection examples, making the happy path plausible. However, the supplied material contains no CI, committed tests, handshake verification, tool contracts, or evidence of controlled error handling. The README and manifest also do not fully agree on tools and configuration, so reproducibility cannot be scored highly.
Security and permissions 8/20: API_TOKEN is marked as secret, and the local example uses an environment variable, which is positive evidence. However, the server accesses the external public web by default and advertises anti-bot access, geo-unblocking, browser automation, and many data tools; the material does not establish least privilege, domain scoping, retention, auditing, or confirmation for risky operations. The hosted example places the token in a URL query string, creating additional exposure risk. No material-supported red-line risk is established, so this is not a 0-4 score.
Maintenance 12/20: The repository is marked unarchived, declares version 2.9.4, uses the MIT license, and provides GitHub issues, documentation, and a support email as maintenance channels. This supports identifiable ownership and an update path. The material does not provide commit cadence, release history, dependency-update evidence, vulnerability-response procedures, or issue-response evidence; stars and issue counts do not add points, so the score remains moderate.
Documentation 15/20: The README covers local installation, hosted access, authentication, modes and tool groups, major tools, configuration, pricing, free-tier limits, examples, and troubleshooting, providing relatively strong surface documentation. Deductions reflect that several claims are not backed by source files or tests, the README and manifest differ on RATE_LIMIT, tool coverage, and defaults, and individual tool parameters, return schemas, permission boundaries, and failure semantics are not sufficiently documented.
Setup experience 14/20: The README provides an npx configuration, Claude Desktop examples, a hosted URL, environment-variable documentation, and troubleshooting for a common runtime error, so mainstream setup is fairly clear. Deductions apply because Node.js and a Bright Data API token are required, the hosted example exposes the token in the URL, platform compatibility and failure diagnosis are limited, and there is no verifiable installation test or CI evidence.

Static review · not runListed 2026-07-29

Read the FMRS scoring method →

Fit and risk

What it can accessUses the networkControls a browser

Best for

AI agents that need live public web data
Research and analytics workflows requiring scraping or structured data
Coding agents that need current npm or PyPI package information
Users requiring browser automation through Pro mode

Not for

Private, unauthorized, or internal data access
Users unwilling to provide a Bright Data API key
Offline environments or local setups without Node.js/npx
Workflows that cannot accommodate usage-based Pro or web-data costs

Required permissions

A Bright Data API_TOKEN is required
The server needs access to Bright Data services and the public web
WEB_UNLOCKER_ZONE and BROWSER_ZONE may be configured optionally
PRO_MODE, GROUPS, TOOLS, timeout, and retry settings may be configured optionally

Risks and side effects

Requests consume account credits; base tools are charged per request and web_data_* tools per returned record
The free tier provides 5,000 credits per month and unused credits do not roll over
PRO_MODE incurs additional usage-based charges
API_TOKEN is a secret credential and should not be exposed in logs or shared configuration
Live web content, availability, and scraping results may depend on target sites and network conditions

Setup

Before you start

Runtime：Node.js · Hosted (no local runtime)

API_TOKEN requiredsecret Bright Data API token, required, obtained from account settings at brightdata.com/cp/setting/users

Other optional settings (9)

WEB_UNLOCKER_ZONE optional Optional custom Web Unlocker zone name, default mcp_unlocker, created in the Bright Data control panel

BROWSER_ZONE optional Optional custom browser zone name, default mcp_browser, created in the Bright Data control panel

PRO_MODE optional Optional boolean; set to true to enable PRO_MODE

GROUPS optional Optional comma-separated tool group IDs (e.g. ecommerce,browser) controlling which tools load

TOOLS optional Optional comma-separated individual tool names enabled on top of groups

POLLING_TIMEOUT optional Optional polling timeout in seconds for web_data_* tools, default 600

BASE_TIMEOUT optional Optional request timeout in seconds for base tools (search and scrape); no limit if unset

BASE_MAX_RETRIES optional Optional max retries for base tools on transient errors (0-3), default 0

RATE_LIMIT optional Optional custom rate limiting, e.g. 100/1h

Configure the server over stdio in an MCP client, run npx @brightdata/mcp, and set API_TOKEN to a Bright Data API key. The README also provides a hosted URL, but that remote option does not change the manifest transport recorded here.

claude_desktop_config.json

{"mcpServers":{"Bright Data":{"command":"npx","args":["@brightdata/mcp"],"env":{"API_TOKEN":"<your-api-token-here>"}}}}

Shown for Claude Desktop. Other clients may use a different file or key (VS Code uses "servers") — the configurator below converts it.

.vscode/mcp.json

{
  "servers": {
    "Bright Data": {
      "command": "npx",
      "args": [
        "@brightdata/mcp"
      ],
      "env": {
        "API_TOKEN": "<your-api-token-here>"
      }
    }
  }
}

Goes in your project's .vscode/mcp.json (VS Code uses a "servers" key).

Terminal

claude mcp add 'Bright Data' -e 'API_TOKEN=<your-api-token-here>' -- npx @brightdata/mcp

Run it in a terminal; replace any <…> placeholders with your own values first.

Generate and validate client config

Check that it works

After connecting, the client tool list should show search_engine, scrape_as_markdown and other tools; run "Search today's weather forecast for New York" to confirm live results come back.

Troubleshooting

For spawn npx ENOENT, check that Node.js is installed or configure the full node path
For complex-site timeouts, check and increase the client timeout
For authentication failures, verify the API_TOKEN and its permissions
For remote connection failures, check network connectivity and firewall settings
For slow data-tool responses, check the POLLING_TIMEOUT setting

Things to try

Once connected, you can ask your AI assistant things like:

Search the web for Tesla's current stock price and give me the source links
Scrape this page as Markdown: https://example.com/article
Get the price and review count for Amazon ASIN B0D2Q9397Y
What's the latest version of express on npm?

Tools 29

search_engine read-only

Runs AI-optimized web search.

search_engine_batch read-only

Runs web searches in batches.

scrape_as_markdown read-only

Converts webpages into AI-ready Markdown.

scrape_batch read-only

Scrapes webpages in batches.

discover read-only

Runs AI-ranked web search with intent-based relevance scoring.

scraping_browser_snapshot read-only

Gets a browser page snapshot.

scraping_browser_click_ref read-only

Clicks a referenced browser element.

scraping_browser_screenshot read-only

Gets a browser page screenshot.

Show 21 more tools

web_data_amazon_product read-only

Extracts structured Amazon product data.

web_data_walmart_product read-only

Extracts structured Walmart product data.

web_data_google_shopping read-only

Extracts Google Shopping data.

web_data_linkedin_posts read-only

Extracts LinkedIn post data.

web_data_tiktok_posts read-only

Extracts TikTok post data.

web_data_youtube_videos read-only

Extracts YouTube video data.

web_data_yahoo_finance_business read-only

Extracts Yahoo Finance business data.

web_data_crunchbase_company read-only

Extracts Crunchbase company data.

web_data_zoominfo_company_profile read-only

Extracts ZoomInfo company profile data.

web_data_zillow_properties_listing read-only

Extracts Zillow property listing data.

web_data_github_repository_file read-only

Extracts GitHub repository file data.

web_data_reuter_news read-only

Extracts Reuters news data.

web_data_google_play_store read-only

Extracts Google Play Store data.

web_data_apple_app_store read-only

Extracts Apple App Store data.

web_data_booking_hotel_listings read-only

Extracts Booking hotel listing data.

web_data_chatgpt_ai_insights read-only

Returns ChatGPT-generated brand insights, citations, and recommendations.

web_data_grok_ai_insights read-only

Returns AI insights from Grok.

web_data_perplexity_ai_insights read-only

Returns search-augmented Perplexity answers with sources.

web_data_npm_package read-only

Returns npm package versions, READMEs, metadata, and dependencies.

web_data_pypi_package read-only

Returns PyPI package information, versions, and project details.

extract read-only

Provides AI-assisted data extraction helpers.

Use cases

Real-time research and fact-checking

E-commerce price, product, and availability analysis

Competitive and market analysis

Web-enabled AI agents

Real-time npm and PyPI package lookup

GEO and brand visibility analysis

Content creation and academic research

Supported clients

Claude Desktop

Claude CodePartial support

CursorPartial support

WindsurfPartial support

Listed from the project's documentation, not tested by this site.

Overview

Bright Data's Web MCP server enables AI agents to search, scrape, extract, and navigate the live public web. It provides Markdown extraction, AI-ranked discovery search, browser automation, and data tools for e-commerce, social, finance, business, research, app stores, travel, GEO, and package intelligence.

Similar servers

Hevy MCP 77 · B

Manage your Hevy workout data from AI assistants via MCP.

★ 517 · Tools 26 Compare with this →

Firecrawl MCP Server 75 · B

Firecrawl's official MCP server for web search, scraping, and structured extraction for AI agents

★ 7.6k · Tools 11 Compare with this →

BioMCP 75 · B

One binary. One grammar. Evidence from the biomedical sources you already trust.

★ 659 · Tools 11 Compare with this →

Compartment 75 · B

Encrypted, fully offline long-term memory for AI agents

★ 583 · Tools 14 Compare with this →

Compare with other MCP servers

Source revision b69d43a1482a Data synced 2026-10-19 Read the FMRS scoring method