CB Insights MCP

How to Install and Use CB Insights MCP

If you want to use the CB Insights MCP, this guide will help you install and set it up easily. CB Insights MCP is a service that connects with the ChatCBI AI agent. You can send messages to ChatCBI and get responses for your projects.

Step 1: Prepare Your Environment

Before starting the installation, make sure you have the right tools and information ready. CB Insights MCP uses uv, a tool to manage the project. You will also need your OAuth Client ID and Secret from CB Insights.

First, install uv by following its official instructions here: uv installation.

You will need to create a file called .env to store important keys. This file contains:

• CBI_CLIENT_ID (your OAuth Client ID)
• CBI_CLIENT_SECRET (your OAuth Client Secret)
• CBI_MCP_TIMEOUT (optional, default is 120 seconds)
• CBI_MCP_PORT (optional, default is 8000)

Make sure to get your Client ID and Secret from the CB Insights API Authentication page.

Step 2: Clone the CB Insights MCP Server

Now you need to get the CB Insights MCP server code onto your computer. Open a terminal and run this command:

git clone https://github.com/cbinsights/cbi-mcp-server.git

This will download the server files to a folder named cbi-mcp-server.

Step 3: Start the MCP Server

Navigate into the downloaded folder:

cd cbi-mcp-server

Run the MCP server using uv with this command:

uv run server.py

This will start the server on your computer, using the port you set in .env or the default port 8000.

Step 4: Use ChatCBI Tool

Once the server is running, you can send messages to ChatCBI through the MCP server. Here is how you send a message:

• Send a request with the key message which is what you want to say.
• Optionally include a chatID if you want to continue a previous conversation.

The MCP server will return a JSON response with fields like:

• chatID — ID of your chat conversation
• message — ChatCBI's response
• relatedContent — List of related info
• sources — Sources used to answer
• suggestions — Suggested follow-up questions
• title — Title of the chat

Step 5: Connect with Claude Desktop (Optional)

If you use Claude Desktop, you can add the MCP server to its configuration file easily. Use the command:

mcp install server.py

This updates the claude_desktop_config.json file and adds:

{
  "mcpServers": {
    "cbi-mcp-server": {
      "command": "/path/to/.local/bin/uv",
      "args": [
        "--directory",
        "/path/to/cloned/cbi-mcp-server",
        "run",
        "server.py"
      ]
    }
  }
}

Make sure to replace the paths with your actual locations.

Step 6: Debugging (If Needed)

If you want to test or find errors, use the MCP inspector tool with this command:

mcp dev server.py

This helps you debug and check your server setup.

---

Now you know how to install and start using the CB Insights MCP server. The steps are simple: set up your environment, download the server, run it with uv, and then you can send messages to ChatCBI. Enjoy building AI agents with CB Insights MCP!