Skip to main content

​
Project Configuration Overview

This guide explains how to set up your project for MCP server deployment. Configuration files should be stored in the root directory of your repository (or in a subdirectory, for monorepos) and define how to build and run your server.

​
Key Concepts

​
Required Configuration Files

  1. Dockerfile: Defines the process for building your server’s container image.
  2. Configuration YAML (e.g., cogni.yaml): Specifies how to start and run your server, including the startup command and exposed port.

​
Automatic Setup

Deployment systems may attempt to auto-generate these files, but manual setup could be necessary if automation fails.

​
File Details

Dockerfile Example Basic Python server container setup: 
FROM python:3.11-slim
WORKDIR /app
COPY . .
RUN pip install -r requirements.txt
CMD ["python", "server.py"]
YAML Configuration Example Specifies how the server should be started: 
version: 1
start:
  command: ["python", "server.py"]
  port: 8000

​
Project Structure Considerations

If using a monorepo or a nested project structure, place the configuration files within the subdirectory containing your server. Be sure to indicate this subdirectory in your server settings during deployment configuration.

​
Best Practices

  • Local Testing: Verify the server builds and runs locally before deployment. Tools like MCP Inspector can assist.
  • Validation: Use a schema to define and validate server configuration options.
  • Efficiency: Optimize Docker images by using minimal base images and multi-stage builds.

​
Deployment

After setting up the configuration files, you can proceed to deploy your server using your chosen deployment method.