Key Takeaways (TL;DR)
- What are MCP Connectors? Model Context Protocol (MCP) servers that let HubSpot Breeze AI agents connect to external tools like Notion, Zapier, Gong, and more — all through natural language.
- Key Benefit: Your Breeze agents can now read, write, and act across your entire tech stack without custom integrations or developer work.
- Cost: Free to install — MCP connector listings carry no additional fees (Breeze Credits and third-party subscriptions may apply).
- Best For: Revenue teams, marketers, and ops leaders who want their AI agents to pull real-time data from project management, analytics, and sales intelligence tools.
- Bottom Line: MCP is the "USB-C for AI" — a single protocol that turns HubSpot Breeze into a true cross-platform AI command center.
➡️ Browse All MCP Connectors for Breeze Agents on the HubSpot Marketplace
If you've ever wished your HubSpot AI agents could reach beyond CRM data to pull insights from Notion, trigger automations in Zapier, or surface competitive intelligence from Gong — that's exactly what just arrived. HubSpot's Spring 2026 Spotlight introduced MCP (Model Context Protocol) support for Breeze Agents, and it fundamentally changes how AI agents interact with your technology stack.
In this post, we'll break down what MCP connectors are, which ones are available today, how they work inside Breeze Studio, and why this matters for every team running on HubSpot.
What Is the Model Context Protocol (MCP)?
Model Context Protocol — or MCP — is an open-source standard originally introduced by Anthropic in late 2024. Often called the "USB-C for AI," MCP creates a standardized, secure, two-way connection between AI applications and the external tools, data sources, and services they need.
Before MCP, connecting an AI agent to five different tools required five custom integrations, each with its own authentication, data format, and maintenance overhead. MCP solves this by providing:
- A universal interface: One protocol that any AI host (like HubSpot Breeze) and any tool provider (like Notion or Gong) can implement.
- Automatic capability discovery: AI agents can dynamically discover what tools and actions are available — no hard-coding required.
- Secure, scoped access: MCP servers expose specific read and write operations, giving teams granular control over what the agent can do.
The architecture follows a clean pattern: MCP Hosts (the AI application, in this case HubSpot Breeze) connect to MCP Servers (provided by each vendor like Notion, Zapier, or Atlassian) through MCP Clients that manage the authentication and communication layer.
How Does MCP Work Inside HubSpot Breeze?
HubSpot's Breeze is the AI platform embedded across every Hub — powering the Prospecting Agent, Customer Agent, Content Agent, and custom agents built in Breeze Studio. With the Spring 2026 update, Breeze now supports MCP natively, meaning you can extend any Breeze Agent with third-party MCP servers.
Setting It Up
The setup process is straightforward:
- Navigate to Breeze > Breeze Studio in your HubSpot account.
- Select (or clone) an existing agent and click Configure agent.
- In the "What this agent can access" section, click Add tool.
- Switch to the MCP Servers tab.
- Hover over the MCP Server you want and click Connect and add.
- Authorize the connection with the third-party provider.
That's it. Your Breeze Agent can now access that tool's data and capabilities through natural language prompts — no code, no API keys, no developer involvement.
Which MCP Connectors Are Available Today?
As of April 2026, HubSpot's Marketplace features eight MCP connectors from major SaaS providers, with more expected as the ecosystem matures.
1. Notion MCP — Documentation & Project Management
- Installs: 200+
- Built by: Notion
- What it does: Allows Breeze Agents to create Notion pages, search across workspaces, manage tasks, and generate documentation — all via natural language.
- Key tools:
notion-search(read-only),notion-fetch(read-only),notion-create-pages(write) - Best for: Marketing teams building campaign briefs, product teams auto-generating PRDs, or anyone who wants their CRM agent to pull project context from Notion.
- Pricing: Free (Breeze Credits may apply)
- View on HubSpot Marketplace →
2. Zapier MCP — 8,000+ App Connections
- Installs: 70+
- Built by: Zapier
- What it does: Connects Breeze Agents to Zapier's entire ecosystem of 8,000+ apps and 30,000+ actions — send messages, create records, schedule events, summarize communications.
- Best for: Teams that need Breeze to interact with tools that don't have their own MCP connector yet. Zapier is the "catch-all" bridge.
- Pricing: Free to install (MCP calls consume 2 Zapier tasks per call; Breeze Credits may apply)
- View on HubSpot Marketplace →
3. Asana MCP — Task & Project Tracking
- Installs: 50+
- Built by: Asana
- What it does: Manage Asana tasks, create project reports and summaries, and surface AI-driven insights from Asana data.
- Best for: Project-heavy teams that want their CRM agent to check task status, assign work, or pull project updates without switching apps.
- Pricing: Free (Breeze Credits may apply)
- View on HubSpot Marketplace →
4. Atlassian MCP — Jira, Confluence & Compass
- Installs: 30+
- Built by: Atlassian
- What it does: Gives Breeze Agents the ability to interact with Jira issues, Confluence pages, and Compass services.
- Best for: Engineering-adjacent teams who need their sales or support agents to check ticket status, reference documentation, or review service health.
- Pricing: Free (Breeze Credits may apply)
- View on HubSpot Marketplace →
5. G2 MCP — Buyer Intent Intelligence
- Installs: 20+
- Built by: G2
- What it does: Arms Breeze Agents with live G2 buyer intent data so they can identify which prospects are actively researching solutions.
- Best for: Sales teams using the Prospecting Agent — layer G2 intent signals on top of HubSpot's native buyer intent for a more complete picture.
- Pricing: Free (Breeze Credits may apply)
- View on HubSpot Marketplace →
6. Linear MCP — Development Workflow
- Installs: 10+
- Built by: Linear
- What it does: Lets Breeze Agents manage issues, projects, and development workflows in Linear.
- Best for: Product-led teams that want to bridge the gap between CRM data and engineering velocity.
- Pricing: Free (Breeze Credits may apply)
- View on HubSpot Marketplace →
7. Amplitude MCP — Product Analytics
- Installs: <10
- Built by: Amplitude
- What it does: Allows Breeze Agents to analyze product data, experiments, and user behavior through natural language queries.
- Best for: Product and growth teams that want AI agents to pull analytics context into CRM conversations — "what features is this account using?"
- Pricing: Free (Breeze Credits may apply)
- View on HubSpot Marketplace →
8. Gong MCP — Revenue Intelligence
- Installs: <10
- Built by: Gong
- What it does: Surfaces AI-powered answers from Gong calls and messages, retrieves deal insights (progress, blockers, objections, next steps), and generates pre-meeting intelligence briefs.
- Key tools:
ask_account(read-only),ask_deal(read-only),generate_brief(read-only) - Best for: Revenue teams who want their Breeze Agents to answer questions like "What objections came up in the last 3 calls with Acme Corp?" without opening Gong.
- Pricing: Free (Breeze Credits may apply)
- View on HubSpot Marketplace →
What Problems Do MCP Connectors Solve?
The Context Gap
The biggest limitation of any CRM-bound AI agent is that it only knows what lives in the CRM. But real business context is scattered across dozens of tools. MCP connectors close this gap by giving Breeze Agents real-time access to:
- Project context (Notion, Asana, Linear, Atlassian)
- Revenue intelligence (Gong, G2)
- Product analytics (Amplitude)
- Thousands of other tools (via Zapier)
The Integration Tax
Traditional integrations require developer time, ongoing maintenance, and periodic break-fix work. MCP connectors install in minutes and follow a standardized protocol that each vendor maintains on their side. Your team doesn't build or maintain the connection.
The Single-Pane-of-Glass Promise
HubSpot has always aimed to be the unified platform for go-to-market teams. MCP connectors move this closer to reality by letting Breeze Agents pull — and push — data across your entire tech stack without users ever leaving HubSpot.
Real-World Use Cases
Sales: Pre-Meeting Intelligence on Autopilot
Imagine a Breeze Agent configured with both Gong MCP and G2 MCP. Before a sales call, the agent can:
- Pull Gong conversation history to surface objections and stakeholder concerns.
- Check G2 buyer intent to see if the prospect is actively evaluating competitors.
- Synthesize a structured briefing document directly in HubSpot.
No tab-switching. No manual research. Just context-rich, AI-generated prep.
Marketing: Campaign Planning Across Tools
A Breeze Agent with Notion MCP and Zapier MCP can:
- Generate a campaign brief in Notion from CRM data.
- Trigger cross-platform workflows via Zapier — posting to Slack, creating Google Calendar events, or updating spreadsheets.
- Track campaign tasks in Asana via the Asana MCP.
Customer Success: Full-Picture Account Reviews
A support-facing Breeze Agent with Amplitude MCP and Gong MCP can:
- Show which product features an account is or isn't using (Amplitude).
- Surface recent conversation themes and risk signals (Gong).
- Combine this with HubSpot ticket and deal data for a complete health check.
Engineering-Sales Alignment
Using Atlassian MCP or Linear MCP, a Breeze Agent can:
- Check the status of Jira issues or Linear projects tied to a customer request.
- Surface Confluence documentation for complex product questions.
- Give sales reps real-time visibility into engineering timelines without Slack back-and-forth.
Implementation Tips
Start Small, Then Expand
Don't connect all eight MCP servers at once. Start with the one or two that address your highest-friction workflows, validate the impact, and then expand.
Clone Before You Customize
HubSpot requires you to clone default Breeze Agents before adding MCP tools. This is actually a best practice — keep your baseline agents clean and build specialized variants for specific use cases.
Mind the Credits
While MCP connectors themselves are free, every interaction consumes Breeze Credits and may trigger usage on the third-party side (for example, each Zapier MCP call uses 2 Zapier tasks). Monitor consumption early to avoid surprises.
Security Considerations
MCP is still a maturing protocol. Earlier in 2026, security researchers identified vulnerabilities in MCP's STDIO transport that could theoretically enable command execution. While HubSpot's implementation uses remote MCP servers (not local STDIO transport), teams in regulated industries should:
- Review each MCP server's permission scope before connecting.
- Limit write access to only the connectors that require it.
- Follow HubSpot's MCP setup guide for configuration best practices.
Pros and Cons
| Pros | Cons |
|---|---|
| Free to install — no upfront cost for connectors | Ecosystem is still small (8 connectors at launch) |
| No-code setup, installs in minutes | Breeze Credits and third-party costs can add up |
| Standardized protocol — vendors maintain their own servers | MCP is a young protocol with evolving security standards |
| Extends Breeze Agents across the full tech stack | Currently works in Breeze Studio (not Breeze Assistant) |
| Read and write scoping gives granular access control | No user reviews yet — early-adopter territory |
| Backed by major vendors (Notion, Zapier, Atlassian, Gong, G2) | Requires HubSpot plans with Breeze Agent access |
How Does This Compare to Salesforce AgentExchange?
It's worth noting that Salesforce's AgentExchange is building a similar ecosystem for Agentforce — pre-built agents and actions from ISV partners. The key difference is the underlying protocol: HubSpot has adopted the open MCP standard, while Salesforce uses its proprietary Agentforce framework.
For organizations running on HubSpot, MCP connectors offer a protocol-native approach that could see broader third-party adoption because the same MCP server that works with Breeze can also work with Claude, Cursor, and other MCP-compatible AI hosts.
What's Coming Next?
HubSpot's developer platform now supports building custom MCP servers, meaning any ISV or internal team can create their own MCP connector for Breeze. The Spring 2026 announcement specifically called out the ability to "Build AI connectors with Remote MCP Server and join an ecosystem alongside Notion, Writer, and more."
Expect this ecosystem to grow rapidly. The companies that build and publish MCP connectors early will become the default integrations for hundreds of thousands of HubSpot portals.
FAQ: MCP Connectors for HubSpot Breeze
What is Model Context Protocol (MCP) and why does HubSpot support it?
MCP is an open-source standard created by Anthropic that provides a universal interface for AI agents to connect to external tools and data sources. HubSpot adopted MCP because it allows any vendor to build a standardized connector that works with Breeze Agents — eliminating the need for custom, one-off integrations.
How much do MCP connectors cost?
All eight MCP connectors currently in the HubSpot Marketplace are free to install. However, using Breeze Agents consumes Breeze Credits, and some connectors (like Zapier MCP, which uses 2 Zapier tasks per call) may trigger usage charges on the third-party side. You'll also need a HubSpot plan that includes Breeze Agent access.
Can I use MCP connectors with Breeze Assistant?
Not currently. MCP connectors are available for agents built and configured in Breeze Studio only. HubSpot has not yet announced MCP support for the general Breeze Assistant experience.
Which MCP connector should I install first?
It depends on your team's workflow. Zapier MCP is the most versatile because it bridges to 8,000+ apps. Notion MCP is ideal if your team already uses Notion for documentation and project management. Gong MCP is a strong choice for revenue teams that want call intelligence baked into their CRM workflows.
Is MCP secure enough for regulated industries?
MCP is a maturing protocol. HubSpot's implementation uses remote MCP servers with scoped permissions, which is more secure than local STDIO transports. However, teams in healthcare, financial services, or other regulated industries should review each connector's permission scope, limit write access where possible, and follow HubSpot's setup guide. Always consult with your security team before enabling new data pathways.
Can I build my own custom MCP connector for Breeze?
Yes. HubSpot's developer platform now supports building and publishing custom MCP servers. If you have a proprietary tool or internal system, you can create an MCP connector that exposes its capabilities to Breeze Agents. Visit HubSpot's developer documentation for details.
How do MCP connectors differ from traditional HubSpot integrations?
Traditional HubSpot integrations (like Data Sync connectors) move structured data between systems on a schedule. MCP connectors give AI agents real-time, on-demand access to external tools through natural language — the agent decides what to query and when, based on the context of the conversation or task it's handling.
MCP connectors represent a fundamental shift in how CRM-embedded AI agents operate. They're no longer limited to what's inside HubSpot — they can reach across your entire tech stack, pull the context they need, and take action.
If you're running on HubSpot and exploring how to maximize your Breeze Agent investments, Vantage Point can help you design, build, and optimize AI agent configurations that leverage MCP connectors for real business outcomes.
➡️ Browse All MCP Connectors for Breeze Agents on the HubSpot Marketplace
Ready to unlock the full potential of HubSpot's AI agents? Contact Vantage Point to discuss your MCP connector strategy and build AI-powered workflows that span your entire tech stack.
