HubSpot

Contacts, companies, deals, and tickets with the HubSpot MCP server in meinGPT

The HubSpot integration connects your assistant directly to your HubSpot CRM, letting you search, read, and edit contacts, companies, deals, and tickets — without leaving meinGPT.

Overview

HubSpot is connected via the official HubSpot MCP server. With the integration you can:

  • Search, read, create, and update contacts, companies, and deals
  • View and manage tickets and their associated activities (calls, emails, meetings, notes, tasks)
  • Retrieve CRM objects by search or by ID
  • Search files in HubSpot

All tool calls run in the context of the signed-in user — your existing HubSpot permissions still apply.

Hinweis

HubSpot used to be connected via a manually created private-app token. The integration now uses OAuth through the HubSpot MCP server. Existing assistants keep the HubSpot connector — you just need to reconnect once.

Requirements

  • A HubSpot account with access to the CRM objects you want to use
  • Your HubSpot user's permissions determine which data the assistant can see and change

Setup

The HubSpot integration uses OAuth — you no longer need to create or store a private-app token manually.

Enable the connector: In the assistant editor, select HubSpot under the connectors and save the assistant.

Connect: Under "Connections" (or on the first tool call in chat), click "Connect". A HubSpot authorization window opens.

Confirm access: Select your HubSpot account and confirm the requested permissions.

Done: After authorization, HubSpot shows as connected under "Connections" and is available in chat.

Available features

Contacts and companies

"Show me the most recently active contacts from HubSpot"
"Create a new contact: Max Mustermann, max.mustermann@example.com, company Example Inc."
"Summarize the latest activity for the company 'Example Inc.'"

Deals

"Which deals in the pipeline are close to closing?"
"Create a new deal for the company 'Example Inc.'"

Tickets

"List all open support tickets with high priority"
"Summarize the latest activity on ticket 12345"

Large knowledge bases: Data Pools

The connector is built for operational, permission-aware queries in the user's context. For company-wide analysis over large datasets, you can additionally ingest HubSpot content into a Data Pool and search it at scale.

  • Connector: permission-aware operational queries per user
  • Data Pool: scalable search and RAG over large datasets
  • Important: the connector's per-user permissions do not automatically apply inside a Data Pool

Learn more under Data Pools (RAG) and Data Sources.

Permissions

When connecting, meinGPT requests read and write access to the core CRM objects:

  • Contacts, companies, deals (read and write) — search, read, create, and update
  • Tickets — tickets and their associated activities (calls, emails, meetings, notes, tasks)
  • Owners and object schemas (read) — to associate records correctly
  • Files (read) — to search HubSpot files

You only see and change data your HubSpot account is allowed to access.

Limitations

  • Access only to objects your HubSpot account is allowed to use
  • Write actions (creating or changing records) happen on your behalf
  • Your HubSpot account's rate limits apply
  • If your HubSpot account has Sensitive Data turned on, the HubSpot MCP server blocks access to activities (calls, emails, meetings, notes, tasks) entirely — regardless of your user's permissions. This restriction is specific to the MCP server and does not apply to the regular HubSpot API.

Privacy and security

  • Authentication uses OAuth 2.1 with PKCE
  • meinGPT stores no HubSpot passwords; the access token is stored encrypted
  • All data is transmitted encrypted
  • You can revoke the connection at any time under "Connections" in meinGPT or in your HubSpot settings
Was this page helpful?