Skip to main content

MCP server integration

Connect AI development tools like Cursor, VS Code, and Claude Desktop to your product documentation through Biel.ai's Model Context Protocol (MCP) server.

Biel.ai provides a hosted Retrieval-Augmented Generation (RAG) layer that makes your documentation searchable and useful to AI tools. This enables smarter completions, accurate technical answers, and context-aware suggestions—directly in your IDE or chat environment.

Biel MCP

The MCP server acts as a bridge between your AI coding assistant and your Biel.ai project, allowing the AI to query your documentation whenever it needs context about your product or API.

Prerequisites​

Before starting, ensure you have:

  • A Biel.ai account. If you don't have one, sign up for free.
  • A project created in your Biel.ai dashboard. Follow the Quickstart guide to create one.
  • Your project slug - this is visible in your project URL or dashboard.
  • An AI tool that supports MCP (Cursor IDE, Claude Desktop, or VS Code with MCP extension).
  • API key (optional) - only required for private projects. Business/Enterprise plans only.

Quickstart​

The fastest way to get started is using our hosted MCP server. No setup or maintenance required.

Use this MCP server configuration with your AI tool:

{
  "mcpServers": {
    "biel-ai": {
      "description": "Query your product's documentation, APIs, and knowledge base.",
      "command": "npx",
      "args": [
        "mcp-remote",
        "https://mcp.biel.ai/sse?project_slug=YOUR_PROJECT_SLUG&domain=https://your-docs-domain.com"
      ]
    }
  }
}

Set the following parameters:

  • project_slug: Your Biel.ai project slug from your dashboard.
  • domain: Your documentation domain.
  • api_key: Optional, API key (only needed for private projects)

Tool-specific setup​

Cursor IDE​

  1. Open Cursor Settings: Ctrl+Shift+J (Windows/Linux) or Cmd+Shift+J (Mac).
  2. Go to MCP → Add new global MCP Server
  3. Paste the configuration from Step 1
  4. Save and restart Cursor.

Claude desktop​

  1. Open your configuration file:
    • macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
    • Windows: %APPDATA%\Claude\claude_desktop_config.json
  2. Add the configuration from Step 1.
  3. Save and restart Claude Desktop.

VS Code​

  1. Install the MCP extension for VS Code.
  2. Open VS Code settings → MCP configuration.
  3. Add the configuration from Step 1.

Advanced: Self-hosted​

For advanced users who prefer to run their own MCP server instance, see the README for self-hosting instructions.

Using the Integration​

Once configured, the MCP server will monitor your AI assistant's messages and automatically route relevant queries to Biel.ai.

Query triggers​

By default, the server routes queries to Biel.ai when:

  • The message includes the keyword biel_ai
  • Or when you've defined custom rules that match the message

You can customize these triggers in your MCP configuration.

Example queries​

Here are some ways to interact with your documentation through the AI tools:

API Documentation​

Can you check in biel_ai what the authentication headers are for the /users endpoint?

Framework Usage​

Could you ask biel_ai how we handle API errors in React?

Deployment guides​

biel_ai, what's our staging deployment checklist?

Troubleshooting​

I'm getting a 403 error from the login API — can biel_ai explain why that might happen?

Setup instructions​

Can you check with biel_ai how to install our plugin in Docusaurus?

Team processes​

Ask biel_ai what new developers need to install before running the app locally.

Configuration options​

The MCP server supports several configuration parameters:

  • project_slug (required): Your Biel.ai project identifier
  • api_key (optional): For authentication with private projects
  • base_url (optional): Defaults to https://app.biel.ai
  • domain (optional): Your documentation domain for additional context
  • chat_uuid (optional): To continue previous conversations

Troubleshooting​

Connection issues​

If your AI tool can't connect to the MCP server:

  1. Verify project slug: Make sure it matches your Biel.ai project.
  2. Network connectivity: Ensure your firewall allows connections to mcp.biel.ai
  3. Check project privacy: For private projects, ensure you have an API key

No responses​

If the AI tool isn't querying Biel.ai:

  1. Include trigger keywords: Make sure to use biel_ai in your messages.
  2. Check project status: Verify your Biel.ai project has indexed content.
  3. Review logs: Check your AI tool's logs for MCP-related errors.

API Key issues​

API keys are only required for private projects and are only available on Business and Enterprise plans.

If you are getting authentication errors:

  1. Check if your project is private: Public projects don't need API keys.
  2. Upgrade your plan: API keys require Business or Enterprise plan.
  3. Regenerate API key: Create a new API key in your team settings if needed.

Security considerations​

  • API keys are passed per request and never stored on the MCP server.
  • Restrict API key permissions to only the necessary Biel.ai projects.
  • Use HTTPS for all communications (enabled by default with hosted service).

Next steps​

Once you have the MCP integration working:

  1. Explore advanced queries: Try asking complex questions about your documentation
  2. Customize trigger rules: Set up custom keywords or patterns for automatic routing.
  3. Monitor usage: Check your Biel.ai analytics to see how the integration is being used.
  4. Train your team: Share example queries and best practices with your development team.

For additional configuration options and advanced usage, check out our API documentation.

Need assistance? Contact us to discuss your specific use case.