HubSpot RevOps MCP Server
A read-only MCP server that connects to HubSpot to audit CRM data, diagnose RevOps maturity, identify at-risk deals, and provide step-by-step guidance for fixing issues.
README
HubSpot RevOps MCP Server
A read-only Model Context Protocol server that connects to your HubSpot account via a Private App and gives an AI assistant (Claude) diagnostic and advisory capabilities over your CRM.
What it does
- Audits contacts, companies, deals, pipelines, and workflows for data quality issues
- Diagnoses RevOps maturity across 4 levels
- Identifies at-risk deals, lifecycle funnel bottlenecks, and critical workflow gaps
- Benchmarks your key sales metrics against industry standards
- Guides you step-by-step on how to fix issues manually in HubSpot
Read-only by design. The MCP server never writes to HubSpot. All changes are made by you in the HubSpot UI.
Setup
1. Create a HubSpot Private App
- In HubSpot: Settings → Integrations → Private Apps → Create a private app
- Give it a name (e.g., "RevOps MCP")
- Under Scopes, enable:
crm.objects.contacts.readcrm.objects.companies.readcrm.objects.deals.readcrm.schemas.contacts.readcrm.schemas.companies.readcrm.schemas.deals.readcrm.objects.owners.readautomation(requires Sales Hub Professional)
- Create the app and copy the Access Token (starts with
pat-na1-…)
2. Clone and install
git clone <your-repo-url>
cd server
npm install
3. Configure environment
cp .env.example .env.local
# Edit .env.local and paste your HubSpot Access Token
4. Run locally
npm run dev
# MCP endpoint: http://localhost:3000/api/mcp
Deploy to Vercel
Option A — Vercel CLI
npm install -g vercel
vercel
# Follow prompts; add HUBSPOT_ACCESS_TOKEN as an environment variable when asked
Option B — Vercel Dashboard
- Push this repo to GitHub
- Import the repo at vercel.com
- Set root directory to
server/ - Add environment variable:
HUBSPOT_ACCESS_TOKEN = pat-na1-… - Deploy
The vercel.json already sets maxDuration: 60 for the MCP route.
Connect to Claude
Once deployed, add the MCP server to Claude Desktop or Claude.ai:
Claude Desktop (claude_desktop_config.json):
{
"mcpServers": {
"hubspot": {
"url": "https://your-deployment.vercel.app/api/mcp"
}
}
}
Claude.ai: Settings → Integrations → Add MCP Server → paste your Vercel URL + /api/mcp
Available Tools (15)
| Tool | What it does |
|---|---|
get_account_overview |
Object counts, pipelines, users |
audit_properties |
Custom + default properties; flags missing critical fields |
audit_pipelines |
Pipeline stages, probabilities, issues |
get_pipeline_health |
Open deals with data quality flags |
get_at_risk_deals |
Deals past close date or with no recent activity |
get_lifecycle_distribution |
Contact funnel counts + conversion rates |
audit_workflows |
All active workflows + stale detection |
check_critical_workflows |
Gap analysis against 12 must-have workflows |
get_deal_metrics |
Win rate, avg deal size, avg cycle, velocity |
get_owner_performance |
Open + closed pipeline per sales rep |
diagnose_revops_maturity |
Level 1–4 maturity score + action plan |
get_how_to_guide |
Step-by-step how-to for any HubSpot task |
search_contacts |
Find contacts by name, email, or company |
get_owners |
List all HubSpot users |
get_common_pitfalls |
Most frequent HubSpot/RevOps mistakes + fixes |
Project structure
server/
├── app/
│ ├── api/
│ │ └── mcp/
│ │ └── route.ts # MCP JSON-RPC 2.0 endpoint
│ └── layout.tsx
├── lib/
│ ├── hubspot.ts # HubSpot API wrapper (read-only)
│ ├── knowledge.ts # Embedded RevOps knowledge base
│ └── tools.ts # Tool definitions + handlers
├── .env.example
├── next.config.ts
├── package.json
├── tsconfig.json
└── vercel.json
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。