home / mcp / cb insights mcp server

CB Insights MCP Server

CB Insights' MCP Server

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "cbinsights-cbi-mcp-server": {
      "command": "/path/to/.local/bin/uv",
      "args": [
        "--directory",
        "/path/to/cloned/cbi-mcp-server",
        "run",
        "server.py"
      ],
      "env": {
        "CBI_MCP_PORT": "8000",
        "CBI_CLIENT_ID": "YOUR_CLIENT_ID",
        "CBI_MCP_TIMEOUT": "120",
        "CBI_CLIENT_SECRET": "YOUR_CLIENT_SECRET"
      }
    }
  }
}

The CB Insights MCP Server provides a pass-through interface to the CB Insights ChatCBI LLM, allowing you to interact with CB Insights through AI agents. It enables seamless messaging, context sharing, and access to related references and sources for assisted decision making.

How to use

You can interact with CB Insights ChatCBI through an MCP client by sending messages and receiving responses. Each message is part of a chat session identified by a chatId. If you provide a chatId, the conversation continues; otherwise a new chat starts.

Key outcomes you can achieve include getting a reply to your question, seeing related content and sources, and receiving suggested follow‑up questions to keep the conversation productive.

How to install

Prerequisites you need before installing include a Python environment and a working MCP setup as described here. You will also use the MCP command line tooling to install and run the server.

Install the CB Insights MCP Server by issuing the installation command from your MCP client or terminal where the MCP tooling is available.

mcp install server.py

# This adds the following MCP configuration for the CB Insights MCP Server:
{
  "mcpServers": {
    "cbi-mcp-server": {
      "command": "/path/to/.local/bin/uv",
      "args": [
        "--directory",
        "/path/to/cloned/cbi-mcp-server",
        "run",
        "server.py"
      ]
    }
  }
}

Additional notes

Configuration and runtime depend on environment variables and the specific MCP tooling flow you choose. The server uses uv to manage the project and relies on OAuth credentials for CB Insights authentication.

Debugging and testing can be performed with the inspector provided by the MCP tooling. Start the server in debug mode to test interactions and inspect messages, responses, and potential issues.

Available tools

ChatCBI

Send a message from an agent to ChatCBI and receive the response, including chatID, related content, sources, and follow-up suggestions.