Docs MCP - Query Bright Data Docs from Your Agent

Connect your AI coding agent to Bright Data’s documentation via Model Context Protocol. Search and retrieve docs as native MCP resources - no scraping required.

What is the Docs MCP?

The Docs MCP server at https://docs.brightdata.com/mcp exposes all Bright Data documentation as searchable MCP resources. Any MCP-compatible coding agent can connect and query the docs in real-time, directly inside its reasoning loop.

Two Bright MCP servers - two different jobs:

MCP Server	URL	Purpose
Docs MCP	https://docs.brightdata.com/mcp	Search and retrieve Bright Data documentation
Web MCP	https://mcp.brightdata.com/mcp?token=<token>	Access live web data - scraping, search, unlocker

This page covers the Docs MCP. To access live web data from your agent, see the MCP Server docs →

---

Connect your agent

Claude Code

claude mcp add --transport http brightdata-docs https://docs.brightdata.com/mcp

Verify the connection:

claude mcp list
# brightdata-docs: https://docs.brightdata.com/mcp (HTTP) - ✓ Connected

Cursor

Open Settings → Tools & Integrations → Add Custom MCP and paste:

{
  "mcpServers": {
    "brightdata-docs": {
      "url": "https://docs.brightdata.com/mcp",
      "transport": "http"
    }
  }
}

VS Code (Copilot)

Add to your .vscode/mcp.json or user MCP settings:

{
  "servers": {
    "brightdata-docs": {
      "url": "https://docs.brightdata.com/mcp",
      "type": "http"
    }
  }
}

Claude Desktop

Add to claude_desktop_config.json (usually at ~/.config/claude/):

{
  "mcpServers": {
    "brightdata-docs": {
      "url": "https://docs.brightdata.com/mcp",
      "transport": "http"
    }
  }
}

Windsurf / Any MCP client

The Docs MCP uses standard HTTP transport - connect any MCP-compatible client to:

https://docs.brightdata.com/mcp

No authentication required. No API token needed.

---

What can your agent do with it?

Once connected, your agent gains a search tool over the full Bright Data documentation index. It can:

• Find the right API - “What’s the parameter for setting a custom geo-location in the SERP API?”
• Retrieve product pages - Pull the full reference for Web Unlocker, Scraping Browser, or any product
• Look up code examples - “Show me a Python example for paginating SERP results”
• Explore integrations - “How do I connect LangChain to Bright Data?”
• Check pricing and limits - “What are the rate limits for the Web MCP free tier?”

The agent queries docs during its reasoning loop - so it always has accurate, up-to-date information rather than relying on stale training data.

---

Use both MCP servers together

For maximum capability, connect both MCP servers in a single session:

# Docs: know how to use Bright Data
claude mcp add --transport http brightdata-docs https://docs.brightdata.com/mcp

# Data: actually use Bright Data
claude mcp add --transport sse brightdata https://mcp.brightdata.com/sse?token=<your-api-token>

Now your agent can look up documentation and execute real web data requests in the same session - no context switching required.

The best setup for a coding agent is all three layers working together:

1. A Skill for embedded baseline knowledge
2. The Docs MCP for on-demand documentation lookup
3. The Web MCP for live web access

Install the skill once, connect both MCP servers, and your agent is fully equipped.

---

Rate limits

Limit	Value
Requests per user per hour	200
Requests per domain per hour	1,000

These limits apply to the Mintlify-hosted Docs MCP. For higher throughput, use llms-full.txt to load the full docs locally.

---

Next steps

Web MCP Server

Add live web data access - search, scrape, and extract from the web

LLM References

Load llms.txt or llms-full.txt for offline or RAG-based access to all docs