Examples
CAMEL-Based Multi-Agent Workflow
This example demonstrates a basic multi-agent system using Coral Protocol and CAMEL agents to collaboratively solve a user query.
Three agents are used:
- Search Agent: Retrieves factual data
- Math Agent: Performs computations
- Interface Agent: Manages interaction with the user
All agents communicate through the Coral Server using the Model Context Protocol (MCP).
Install the Python dependencies:
Install CAMEL with all optional dependencies:
Troubleshooting
- Agents are limited to 20 iterations to manage token usage. Restart agents if needed.
- Agents do not automatically unregister. Restart the server before re-running the system.
- If anything breaks or seems unclear, reach out on Discord: Join the community
Build on This Example
To expand your agent society:
- Duplicate any existing agent script
- Assign a unique role and agent name
- Register it with the Coral Server
Each new agent can plug into the same thread-based architecture.
Future Potential
This example serves as a proof of concept. As the protocol matures, features like remote-mode execution, session isolation, and lifecycle management will extend its robustness.
We welcome experimentation and contributions as the ecosystem grows.