Hosting

Remote hosted or local self-hosted

Transport

SSE or Streamable HTTP (remote); stdio (local)

Endpoint

https://mcp.brightdata.com/mcp?token={BRIGHT_DATA_API_TOKEN} (Streamable HTTP); https://mcp.brightdata.com/sse?token={BRIGHT_DATA_API_TOKEN} (SSE)

Authentication

Bright Data API token. Remote setup passes the token as the token query parameter; local setup uses API_TOKEN.

Requirements

• Bright Data account and API token
• MCP-compatible client
• Local mode requires Node.js/npm; remote mode requires no local server runtime

Setup steps

1. For hosted use, add the documented Streamable HTTP or SSE endpoint and replace the token placeholder through the client's secure secret mechanism.
2. For local use, configure npx @brightdata/mcp with API_TOKEN set locally; PRO_MODE and GROUPS are optional.
3. Choose the needed mode/tool groups, complete client authentication/configuration, then reload and verify with a harmless web search.

All 32 official tool names are now represented in detailed tool rows.

• search_engine

  Search Google, Bing, or Yandex and return search results.

  01
• scrape_as_markdown

  Fetch a webpage and return extracted Markdown.

  02
• search_engine_batch

  Run up to ten search queries in parallel.

  03
• scrape_batch

  Fetch up to ten webpages and return Markdown.

  04
• discover

  Search and rank web results for an intent.

  05
• scrape_as_html

  Fetch a webpage and return HTML.

  06
• extract

  Extract structured JSON from a webpage using the documented extraction feature.

  07
• session_stats

  Report tool calls for the current session.

  08
• search_scrapers

  Find a prebuilt scraper by site or data type.

  09
• get_scraper_details

  Return a scraper description, input fields, sample input, and output fields.

  10
• run_scraper

  Run a scraper method on supplied inputs.

  11
• get_scraper_progress

  Get the status of a scraper run.

  12
• get_scraper_results

  Retrieve results from a scraper run.

  13
• refresh_scrapers

  Refresh the prebuilt scraper list.

  14
• list_dataset_fields

  List filterable fields for a searchable dataset.

  15
• search_dataset

  Search a Bright Data dataset with filters and pagination.

  16
• scraping_browser_navigate

  Navigate the active scraping-browser session to a URL.

  17
• scraping_browser_go_back

  Navigate the active scraping-browser session back.

  18
• scraping_browser_go_forward

  Navigate the active scraping-browser session forward.

  19
• scraping_browser_snapshot

  Capture an accessibility snapshot of the active browser page.

  20
• scraping_browser_type_ref

  Type text into an element identified by a reference from the latest ARIA snapshot.

  21
• scraping_browser_fill_form

  Fill multiple browser form fields identified by ARIA references.

  22
• scraping_browser_screenshot

  Capture a screenshot of the current browser page.

  23
• scraping_browser_network_requests

  List network requests recorded by the current browser session.

  24
• scraping_browser_wait_for_ref

  Wait for a referenced page element to become visible.

  25
• scraping_browser_get_text

  Return text content from the current page body.

  26
• scraping_browser_get_html

  Return HTML content from the current browser page.

  27
• scraping_browser_scroll

  Scroll the current browser page to its bottom.

  28
• scraping_browser_scroll_to_ref

  Scroll until a referenced page element is in view.

  29
• chatgpt_ai_insights

  Send a prompt to ChatGPT and return its answer as Markdown.

  30
• perplexity_ai_insights

  Send a prompt to Perplexity and return its answer as Markdown.

  31
• scraping_browser_click_ref

  Click a page element using its reference from the latest ARIA snapshot.

  32

• The 5,000 free requests/month figure is account-level and shared across team users according to current docs; account plan/usage may change.
• Rapid, Pro, and tool-group modes expose different capabilities.
• API token must be kept secret; remote setup embeds it in an endpoint query parameter, so use client secret storage and do not share the resulting URL.

• official tool reference

  Reviewed current publisher or maintainer documentation for the tool inventory, availability, endpoint, or configuration.

• official documentation

  Reviewed current publisher or maintainer documentation for the tool inventory, availability, endpoint, or configuration.

• official vendor documentation

  Official docs enumerate all 32 server tools, all enabled by default; the detailed catalog now stores all 32 names.

Documentation checked 2026-10-28T00:00:00Z. Connection has not been runtime tested.