Welcome to MCP-CRM. This is a modern CRM (Customer Relationship Management) system. Its main feature is a built-in MCP (Model Context Protocol) Connector. This connector allows an AI Agent to control the database directly.
- Built-in MCP Connector: AI Agents can connect to the CRM backend and manage data safely using specific tools (like
create_schema,add_record,query_data). - Flexible Data Model: Build custom objects for any use case. The data model adapts to your business, not the other way around. Get relationship context with automatic enrichment, and create associations for accurate, actionable insights.
- Secure Authentication: Secure user login powered by Firebase.
This project uses a Monorepo architecture powered by Turborepo.
- Hosting & Backend: Cloudflare Workers.
- Database: Cloudflare D1 (Serverless SQLite).
- Frontend: React inside Turborepo.
- Styling: Vanilla CSS / CSS Modules.
- Authentication: Firebase Auth.
├── apps/
│ ├── crm-website/ # Public-facing, SEO-optimized Next.js app
│ ├── crm-client/ # The custom UI (CRM dashboard in React SPA)
│ ├── mcp-server/ # The backend running on Cloudflare Workers
│ └── storybook/ # Isolated Storybook environment for UI components
├── packages/
│ ├── ui/ # The Design System package (components using Vanilla CSS)
│ ├── tsconfig/ # Shared TypeScript configurations
│ └── eslint-config/ # Linter rules (main defense against AI mistakes)
├── package.json
└── turbo.json
- User Interface (
crm-client&crm-website): The user interacts with the Next.js website and logs into the React dashboard using Firebase. - Backend Server (
mcp-server): The dashboard sends API requests to the Cloudflare Worker backend. - Flexible Database (Cloudflare D1): Data is stored in Cloudflare D1. The system uses a flexible schema pattern so you can build custom objects and relationships on the fly.
- AI Connection (MCP): The AI Agent connects to the
mcp-server. It uses strict explicit tools to safely read, write, and update the data model.