Bright Data MCP server
Bright Data's official server gives agents reliable web search and scraping that gets past blocks, CAPTCHAs, and geo-restrictions.
Bright Data MCP is the official server from Bright Data, a large web-data platform whose proxy network and Web Unlocker are built to retrieve public pages that ordinary fetchers get blocked from. The default rapid toolkit gives an agent five always-enabled tools: search the web with AI-optimized results, run batched searches, scrape any page to clean markdown, scrape pages in batches, and discover sources via AI-ranked, intent-based relevance scoring. Because requests route through Bright Data's unblocking infrastructure, the server reliably returns content from sites that throttle bots, present CAPTCHAs, or vary by geography.
Setting PRO_MODE=true unlocks 60+ additional tools — structured extractors for e-commerce, social, and other domains plus a remote browser for interactive sessions — and GROUPS or TOOLS let you whitelist just the bundles you need. It runs locally over stdio via npx @brightdata/mcp with an API_TOKEN, and Bright Data also offers a hosted remote endpoint for zero-setup use. The free tier includes 5,000 requests per month.
Quick install
Copy-paste configs are provided for all 8 supported clients. Pick your client below.
Add to ~/.claude.json
{
"mcpServers": {
"bright-data": {
"type": "http",
"url": "https://mcp.brightdata.com/mcp",
"headers": {
"Authorization": "Bearer <API_TOKEN>"
}
}
}
}claude mcp add --transport http bright-data https://mcp.brightdata.com/mcpAvailable tools
| Tool | Description |
|---|---|
| search_engine | Web search with AI-optimized results for research, fact-checking, and current events (base tool, always enabled). |
| search_engine_batch | Runs multiple web searches in a single batched call (base tool, always enabled). |
| scrape_as_markdown | Converts any webpage to clean markdown, getting past blocks and CAPTCHAs (base tool, always enabled). |
| scrape_batch | Scrapes multiple pages to markdown in a single batched call (base tool, always enabled). |
| discover | AI-ranked web search with intent-based relevance scoring for deep research and RAG pipelines (base tool, always enabled). |
Required configuration
- API_TOKENRequired
Bright Data API token used to authenticate requests.
- PRO_MODEOptional
Set to true to enable all 60+ tools beyond the base toolkit.
- RATE_LIMITOptional
Optional request throttle, e.g. 100/1h.
- WEB_UNLOCKER_ZONEOptional
Optional custom Web Unlocker zone name.
- BROWSER_ZONEOptional
Optional custom Browser zone name for interactive sessions.
What you can do with it
Scrape sites that block bots
An agent pulls clean markdown from pages behind CAPTCHAs, rate limits, or geo-walls because requests route through Bright Data's Web Unlocker network.
Deep research at volume
Use discover and batched search to gather many sources quickly, then scrape_batch to retrieve their full content for a RAG pipeline in one pass.
FAQ
- Is it free?
- The server is open source and includes a free tier of 5,000 requests per month; higher volume and Pro-mode tooling bill against your Bright Data account.
- Does it support remote/OAuth?
- It runs locally over stdio via npx @brightdata/mcp with an API_TOKEN, and Bright Data also offers a hosted remote endpoint authenticated by the same token rather than OAuth.