Built something for scaling MCP apps - would love feedback

[dhyansraj]

Pre-submission Checklist

• I have verified that this discussion would not be more appropriate as an issue in a specific repository
• I have searched existing discussions to avoid duplicates

Discussion Topic

Hi MCP community! 👋

I wanted to share a project I've been working on that extends MCP to distributed environments: MCP Mesh.

What it does:
MCP Mesh lets you scale MCP applications across Kubernetes with the same simplicity as running them locally. The core idea is dynamic dependency injection - agents declare what they need with @mesh.tool decorators, and the runtime automatically wires them together.

Key features:

• Agents work standalone but enhance when dependencies become available
• Remote MCP calls look identical to local function calls
• Kubernetes-native with auto-scaling and fault tolerance
• Full MCP protocol compatibility

Example:
@mesh.tool(dependencies=["date_service"])
def hello_mesh(date_service=None):
if date_service:
return f"Hello! Today is {date_service()}"
return "Hello! (Date service not available)"

The date_service parameter gets automatically injected with a proxy to another agent's function - no networking code required.

Links:

• GitHub: https://github.com/dhyansraj/mcp-mesh
• Quick install: pip install mcp-mesh

I'd love to hear your thoughts and feedback! The goal is to make it easier for the MCP ecosystem to scale to production environments while keeping the same elegant simplicity MCP is known for.