🚀 Live Project
🌐 Live Demo: https://open-relay.vercel.app
📹 Demo Video: https://youtu.be/bTqQ6mbHEyQ
💻 GitHub Repository: https://github.com/gdhanushkumar07/OpenRelay
Inspiration
As AI agents become more capable, enterprises will increasingly rely on them to automate business operations. However, organizations still communicate through tightly coupled APIs, custom integrations, and isolated systems.
We asked a simple question:
What if AI agents from different organizations could securely communicate through a common protocol instead of every company building integrations with every other company?
That idea became OpenRelay—a secure coordination layer powered by the AICOO Protocol.
To demonstrate the concept, we built an enterprise e-commerce procurement workflow where AI agents from multiple organizations collaborate to complete a customer order.
What it does
OpenRelay enables secure communication and coordination between AI agents across different organizations.
Instead of direct organization-to-organization integrations, every request flows through OpenRelay using the AICOO Protocol.
For our demo, a customer purchases a product from a marketplace.
OpenRelay securely coordinates multiple enterprise AI agents:
- Marketplace Agent
- Supplier Agent
- Warehouse Agent
- Finance Agent
- Insurance Agent
- Courier Agent
Each agent performs only its assigned responsibility while OpenRelay manages:
- Secure request routing
- Identity verification
- Permission validation
- Session management
- Workflow orchestration
- Context isolation
- Response aggregation
The result is a secure and standardized communication layer for enterprise AI systems.
How we built it
Frontend
- Next.js 16
- React
- TypeScript
- Tailwind CSS
Architecture
We designed OpenRelay around the AICOO Protocol, which acts as the communication layer between enterprise AI agents.
The application demonstrates:
- Multi-agent coordination
- Enterprise workflow orchestration
- Secure request routing
- Cross-organization communication
- Customer and seller perspectives
- Interactive workflow visualization
- Live procurement dashboard
Challenges we ran into
One of the biggest challenges was designing a workflow that felt like a real enterprise system rather than a simple AI chatbot.
We spent significant time:
- Designing realistic multi-organization workflows
- Visualizing AI agent communication
- Creating customer and seller perspectives
- Building interactive workflow timelines
- Making complex coordination easy to understand
Balancing enterprise-level concepts with an intuitive user experience was our biggest design challenge.
Accomplishments that we're proud of
- Built a complete enterprise AI coordination platform from scratch.
- Designed an intuitive visualization of cross-organization AI communication.
- Created a realistic enterprise procurement workflow involving multiple organizations.
- Developed a clean and interactive UI suitable for technical demonstrations.
- Successfully demonstrated how the AICOO Protocol can coordinate multiple AI agents securely.
What we learned
Throughout this project we learned that enterprise AI is not only about building smarter AI agents.
The real challenge is enabling AI systems from different organizations to collaborate securely.
We also gained experience in:
- Multi-agent system design
- Enterprise workflow orchestration
- Secure communication architecture
- Interactive system visualization
- Building scalable frontend applications with Next.js
What's next for OpenRelay
This project is an early demonstration of protocol-level AI coordination.
Future plans include:
- Native AICOO Protocol integration
- Live enterprise API connectivity
- Persistent workflow execution
- Organization onboarding
- Policy-based permission management
- Distributed audit logging
- Real-time event streaming
- AI-to-AI negotiation between organizations
- Support for additional enterprise workflows beyond procurement
Our long-term vision is for OpenRelay to become the secure communication layer that enables AI agents from different organizations to work together seamlessly.
Built With
- nextjs
- react
- tailwind
- typescript
Log in or sign up for Devpost to join the conversation.