Welcome to Notify-MCP¶
Seamless Team Collaboration Across AI Platforms¶
Notify-MCP is a powerful pub-sub MCP server that enables teams to share notifications, decisions, and status updates across different AI assistants (Claude, ChatGPT, Gemini).
Why Notify-MCP?¶
-
Cross-Platform Collaboration
Share information seamlessly between Claude, ChatGPT, Gemini, and other AI assistants. Your team stays synchronized regardless of which platform they use.
-
Persistent Storage
SQLite-based persistent storage enables true team collaboration. Share a database file and everyone sees the same channels and notifications.
-
Smart Filtering
Subscribe to channels with intelligent filters based on priority, tags, themes, and sender roles. Only receive what matters to you.
-
Production Ready
62 passing tests, 70% code coverage, comprehensive documentation. Built with SQLAlchemy, aiosqlite, and modern async Python.
Quick Example¶
1. Developer publishes architecture decision in Claude¶
# In Claude Code or Claude Desktop
"Create a channel called 'architecture' and publish a decision about
migrating to microservices architecture"
2. Team members receive notification¶
All team members subscribed to the architecture channel—whether using Claude, ChatGPT, or Gemini—see the decision:
{
"title": "Migration to Microservices",
"body": "Decision: Moving from monolith to microservices using Docker/K8s",
"priority": "high",
"theme": "architecture-decision",
"sender": "Alice (Developer)",
"timestamp": "2025-01-12T10:30:00Z"
}
3. Everyone stays aligned¶
No more information silos! Architecture decisions, incidents, and project updates flow seamlessly across your entire team.
Key Features¶
🎯 Multi-Channel System¶
Create dedicated channels for teams, projects, or topics. Organize notifications logically.
🔔 Pub-Sub Architecture¶
Decoupled notification delivery. Publishers and subscribers don't need to know about each other.
💾 Persistent Storage¶
Version 1.1.0+ includes SQLite storage for team collaboration. Notifications survive server restarts.
📊 Rich Notification Model¶
- Sender info: User ID, name, role, AI tool
- Context: Theme, priority, tags, validity
- Information: Title, body (text/markdown/json)
- Actions: Optional action buttons
- Metadata: ID, timestamp, sequence, threading
🔐 Type-Safe¶
Full Pydantic validation with JSON Schema. Catch errors early.
🧪 Well-Tested¶
62 unit tests, 70% code coverage, comprehensive test suite.
Use Cases¶
-
Architecture Decisions
Document and broadcast architectural decisions to all stakeholders. Maintain a searchable history of technical choices.
-
DevOps Incidents
Alert teams about production incidents. Critical notifications reach everyone instantly, regardless of their AI platform.
-
Team Coordination
Coordinate work across distributed teams. Share status updates, milestones, and blockers in real-time.
-
Project Updates
Broadcast project milestones, sprint updates, and requirement changes. Keep stakeholders informed automatically.
Getting Started¶
Ready to enable seamless collaboration across your team's AI assistants?
- Install Notify-MCP - Set up in 5 minutes
- Configure Storage - Enable team collaboration
- Create Channels - Organize your notifications
- Explore Use Cases - Learn real-world applications
Community & Support¶
GitHub
Find the source code, report issues, and contribute on GitHub.
Documentation
Comprehensive guides, API reference, and examples available throughout this site.
Discussions
Join the conversation in GitHub Discussions.
Bug Reports
Found a bug? Create an issue on GitHub.
What's New¶
Version 1.1.0 - Persistent Storage 🎉¶
The latest release introduces SQLite-based persistent storage, enabling true team collaboration:
- ✅ Persistent across restarts - Notifications survive server restarts
- ✅ Team collaboration - Share database files for cross-team sync
- ✅ Zero setup - File-based SQLite requires no server
- ✅ Production ready - 62 tests, 70% coverage
Built with ❤️ for seamless AI collaboration
Made with MCP (Model Context Protocol) by Anthropic