HubSpot MCP Server: Give Your AI Agent Full CRM Access

Why HubSpot Needs an MCP Server

HubSpot is where sales, marketing, and support teams live. Contacts, deals, tickets, email sequences, landing pages, reports. When your AI agent can read and write HubSpot directly, it handles the CRM busywork that eats hours every week.

The HubSpot API requires OAuth app registration, scope management, rate limit handling, and association API knowledge. DataFaucet captures it all from the web UI in 60 seconds. No developer portal, no app approval process.

What You Get

After scanning HubSpot, DataFaucet generates tools like:

• search_contacts — find contacts by name, email, company, or custom properties
• get_deal — fetch deal details including stage, amount, and associated contacts
• update_deal_stage — move deals through your pipeline
• create_contact — add new contacts with properties
• list_tickets — query support tickets by status or priority
• get_company — fetch company records with associated deals and contacts
• search_emails — find logged email activities

Each tool has typed parameters matching HubSpot's property system.

How to Create It

Step 1: Open DataFaucet

Go to datafaucet.dev and choose the Browse tab.

Step 2: Navigate to HubSpot

Paste https://app.hubspot.com and log into your portal. DataFaucet captures your session automatically.

Step 3: Interact Naturally

Open a few contacts, view a deal, search for a company, check a ticket, browse a report. DataFaucet captures every API call the HubSpot frontend makes.

Step 4: Deploy

Click Deploy. Your MCP server is live with an SSE endpoint ready for any MCP client.

Example: AI-Powered Deal Review

With a HubSpot MCP server connected to Claude:

You: "Show me all deals in the Negotiation stage over $25k that haven't been touched in a week"
Claude: [calls search_deals with stage and last-modified filters]
Found 5 deals totaling $187,500:
- Acme Corp renewal ($45,000) — last activity 9 days ago
- DataFlow Pro ($32,000) — last activity 12 days ago
...
Want me to draft follow-up tasks for each?

No SOQL. No report builder. No manual CRM searching.

Why Not Use the HubSpot API Directly?

Building a custom MCP server against the HubSpot API requires:

1. Create a developer app in HubSpot's developer portal
2. Configure OAuth 2.0 with required scopes (crm.objects.contacts.read, etc.)
3. Implement token refresh (tokens expire every 30 minutes)
4. Learn the CRM v3 API and associations model
5. Handle rate limiting (100 requests per 10 seconds)
6. Implement search API with filter groups syntax
7. Deploy, host, and maintain the server

DataFaucet bypasses all of this. It captures the exact requests the HubSpot frontend makes.

Common Use Cases

• Sales reps — AI preps for calls by pulling contact history and deal context
• Sales managers — AI generates pipeline reports and flags stale deals
• Marketing — AI checks campaign performance, creates contact lists
• Support — AI pulls ticket context, suggests responses based on history
• RevOps — AI audits data quality, finds duplicate contacts, checks property completeness

Works With

• Claude Desktop (claude_desktop_config.json)
• Cursor (MCP settings)
• Windsurf (MCP panel)
• Codex CLI
• Any MCP-compatible client

Get Started

Create your HubSpot MCP server in 60 seconds. Free tier, no credit card, 3 servers included.

Also see: Zendesk MCP Server for connecting AI to customer support workflows, or Salesforce MCP Server for full CRM access.