This repository provides a Model Context Protocol (MCP) server for integrating AI models with HubSpot CRM, enabling interaction with HubSpot data through tools like contact management. It allows AI to
This is a Model Context Protocol (MCP) server designed to integrate AI models with HubSpot CRM, enabling interaction with HubSpot data through a standardized interface. It leverages Anthropic's MCP for seamless operation.
No resources are implemented. Tools are the primary focus.
Demonstrates example prompts for creating HubSpot contacts from LinkedIn profiles and retrieving company pipeline activities.
The server provides tools for managing HubSpot objects:
hubspot_create_contact
creates contacts, checking for duplicates based on name and company.hubspot_create_company
creates companies, checking for duplicates based on name, and hubspot_get_company_activity
retrieves company activity history.hubspot_get_recent_engagements
retrieves recent activities, hubspot_get_recent_companies
retrieves recently active companies, and hubspot_get_recent_contacts
retrieves recently active contacts.Automatic installation for Claude Desktop via Smithery is supported.
Requires a HubSpot access token, obtainable by creating a private app with specific scopes (tickets, CRM object read/write permissions, sales email read, etc.).
Supports pulling from Docker Hub or building locally. Multi-platform publishing is supported via docker buildx
.
Automatic installation for Claude Desktop via Smithery is supported.
Provides JSON configurations for running the server with Docker, including passing the access token as an environment variable or command-line argument.
Development environment setup instructions are provided.
Licensed under the MIT License.
baryhuang/mcp-hubspot
December 27, 2024
March 28, 2025
Python