playbooks

home / mcp / hubspot mcp server

HubSpot MCP Server

Provides an MCP server to connect Claude and other AI assistants to HubSpot CRM data with vector-backed search and caching.

Installation
Add the following to your MCP client configuration file.

Configuration

View docs
{
  "mcpServers": {
    "baryhuang-mcp-hubspot": {
      "command": "docker",
      "args": [
        "run",
        "-i",
        "--rm",
        "-e",
        "HUBSPOT_ACCESS_TOKEN=your_token",
        "-v",
        "/path/to/storage:/storage",
        "buryhuang/mcp-hubspot:latest"
      ],
      "env": {
        "HUBSPOT_ACCESS_TOKEN": "your_token"
      }
    }
  }
}

This MCP server lets AI assistants securely access HubSpot CRM data, including contacts, companies, and engagement details, while providing fast semantic search and reliable caching to handle HubSpot API limits and multi-step workflows.

How to use

You connect your AI assistant to HubSpot data through an MCP client, enabling natural language interactions with contacts, companies, and engagement history. The server stores conversation threads, uses semantic search to surface relevant past interactions, and caches embeddings for fast responses. Use this to build AI-enabled workflows like creating contacts from profiles, retrieving recent activity, or querying data across your HubSpot account.

How to install

Prerequisites: you need a HubSpot access token with the following scopes: crm.objects.contacts (read/write), crm.objects.companies (read/write), and sales-email-read.

Install via Smithery

# Install via Smithery (recommended)
npx -y @smithery/cli@latest install mcp-hubspot --client claude

Run via Docker

docker run -e HUBSPOT_ACCESS_TOKEN=your_token buryhuang/mcp-hubspot:latest

Claude desktop configuration

{
  "mcpServers": {
    "hubspot": {
      "command": "docker",
      "args": [
        "run", "-i", "--rm",
        "-e", "HUBSPOT_ACCESS_TOKEN=your_token",
        "-v", "/path/to/storage:/storage",
        "buryhuang/mcp-hubspot:latest"
      ]
    }
  }
}

What to expect after setup

Once running, you can prompt your AI assistant to perform HubSpot operations such as creating contacts or companies with duplicate prevention, retrieving recent activity, or performing semantic searches over retrieved data. The server persists data to your configured storage and maintains thread-level indexes for precise retrieval across conversations.

Additional notes

The server is designed for zero-configuration deployment with Docker, while also supporting local development and multi-platform Docker images for broad compatibility.

Available tools

hubspot_create_contact

Create a HubSpot contact with duplicate prevention to avoid creating multiple similar entries.

hubspot_create_company

Create a HubSpot company with duplicate prevention to avoid duplicate company records.

hubspot_get_company_activity

Retrieve activity for a specific HubSpot company, including emails, calls, and meetings.

hubspot_get_active_companies

Retrieve the most recently active companies based on recent interactions and activity.

hubspot_get_active_contacts

Retrieve the most recently active contacts based on recent interactions and activity.

hubspot_get_recent_conversations

Fetch recent conversation threads with messages for context retention.

hubspot_search_data

Perform semantic search across previously retrieved HubSpot data using vector embeddings.