feat: change MCP's default port from 5000 to 8000 (#85)

Author: DaleSeo

add-mcp/.apollo/mcp.local.yaml

@@ -20,7 +20,7 @@ introspection:
     enabled: true
 transport:
   type: streamable_http
-  port: 5050                    # MCP server port
+  port: 8000                    # MCP server port (default: 8000)
 endpoint: http://localhost:4000 # Your GraphQL API endpoint
 
 # =============================================================================
@@ -40,7 +40,7 @@ description: "GraphQL API for ${PROJECT_NAME}"
 # 1. Create operation collections in Apollo Studio Sandbox
 # 2. Update endpoint to match your GraphQL server URL
 # 3. Run: docker build -f mcp.Dockerfile -t my-mcp-server .
-# 4. Run: docker run -p 5050:5050 --env-file .env my-mcp-server
+# 4. Run: docker run -p 8000:8000 --env-file .env my-mcp-server
 
 # Environment variables (optional):
 # GRAPHQL_ENDPOINT - override endpoint URL

add-mcp/.apollo/mcp.staging.yaml

@@ -20,7 +20,7 @@ introspection:
     enabled: true
 transport:
   type: streamable_http
-  port: 5050                    # MCP server port
+  port: 8000                    # MCP server port (default: 8000)
 endpoint: "${STAGING_GRAPHQL_ENDPOINT}"  # Your staging GraphQL API endpoint
 
 # Authentication headers for staging
@@ -53,7 +53,7 @@ description: "Staging environment for ${PROJECT_NAME} GraphQL API"
 # 2. Set STAGING_GRAPHQL_ENDPOINT to your staging GraphQL URL
 # 3. Set APOLLO_KEY for authentication
 # 4. Run: docker build -f mcp.Dockerfile -t my-mcp-staging .
-# 5. Run: docker run -p 5050:5050 --env-file .env.staging my-mcp-staging
+# 5. Run: docker run -p 8000:8000 --env-file .env.staging my-mcp-staging
 
 # Environment variables for staging:
 # STAGING_GRAPHQL_ENDPOINT - staging GraphQL endpoint

add-mcp/.env.template

@@ -22,7 +22,7 @@ APOLLO_GRAPH_REF="{{APOLLO_GRAPH_REF}}"
 
 # MCP Server Configuration
 MCP_ENABLE=1
-MCP_PORT=5050
+MCP_PORT=8000
 
 # Optional: Environment and logging
 NODE_ENV="development"

add-mcp/MCP_README.md

@@ -21,7 +21,7 @@ All through natural conversation, using your existing GraphQL API.
 Your MCP server details:
 ```
 Server name: {{PROJECT_NAME}}
-MCP endpoint: http://127.0.0.1:5050/mcp
+MCP endpoint: http://127.0.0.1:8000/mcp
 ```
 
 For Claude Desktop setup instructions:
@@ -48,7 +48,7 @@ set -a && source .env && set +a && rover dev --supergraph-config supergraph.yaml
 
 # This starts:
 # → GraphQL API: http://localhost:4000
-# → MCP Server: http://localhost:5050
+# → MCP Server: http://localhost:8000
 ```
 
 **Done!** In Claude Desktop, look for your MCP server named `{{PROJECT_NAME}}` in the available tools. Your GraphQL API is now accessible to your AI assistant!
@@ -84,13 +84,13 @@ If you prefer to run just the MCP server separately (without rover dev):
 ```bash
 # Build and run with Docker
 docker build -f mcp.Dockerfile -t your-project-mcp .
-docker run -d --name your-project-mcp -p 5050:5050 --env-file .env your-project-mcp
+docker run -d --name your-project-mcp -p 8000:8000 --env-file .env your-project-mcp
 
 # Verify it's running
-curl http://localhost:5050/health
+curl http://localhost:8000/health
 
 # Test with MCP inspector
-npx @modelcontextprotocol/inspector --transport http --server-url http://localhost:5050/mcp
+npx @modelcontextprotocol/inspector --transport http --server-url http://localhost:8000/mcp
 ```
 
 ## Prerequisites
@@ -123,6 +123,6 @@ npx @modelcontextprotocol/inspector --transport http --server-url http://localho
 ## Need Help?
 
 - **AI client not connecting?** Ensure you've followed the setup instructions from `rover docs open mcp-qs` and restarted your AI client completely.
-- **Port conflicts?** Rover dev uses ports 4000 (GraphQL) and 5050 (MCP). Check nothing else is using these ports.
+- **Port conflicts?** Rover dev uses ports 4000 (GraphQL) and 8000 (MCP). Check nothing else is using these ports.
 - **Environment variables not loading?** Make sure you're using `set -a && source .env && set +a` to properly export variables before running `rover dev`.
 - **Need more help with MCP server and tools?** Visit our [Apollo MCP server troubleshooting guide](https://www.apollographql.com/docs/apollo-mcp-server/quickstart#troubleshooting).

add-mcp/mcp.Dockerfile

@@ -14,8 +14,8 @@
 #
 # USAGE:
 # 1. Build: docker build -f mcp.Dockerfile -t {{PROJECT_NAME}} .
-# 2. Run:   docker run -p 5050:5050 --env-file .env {{PROJECT_NAME}}
-# 3. Test:  npx @mcp/inspector (Transport: HTTP, URL: http://localhost:5000/mcp)
+# 2. Run:   docker run -p 8000:8000 --env-file .env {{PROJECT_NAME}}
+# 3. Test:  npx @mcp/inspector (Transport: HTTP, URL: http://localhost:8000/mcp)
 #
 # =============================================================================
 
@@ -51,7 +51,7 @@ COPY .apollo/mcp.local.yaml /mcp.yaml
 
 # Expose the MCP server port
 # The MCP server listens on this port for connections from AI assistants
-EXPOSE 5050
+EXPOSE 8000
 
 # =============================================================================
 # STARTUP COMMAND
@@ -62,7 +62,7 @@ EXPOSE 5050
 # 1. Auto-discover configuration files in current directory
 # 2. Load tools from MCP_TOOLS_DIR (/app/tools)
 # 3. Connect to GRAPHQL_ENDPOINT
-# 4. Start MCP server on MCP_PORT (5050)
+# 4. Start MCP server on MCP_PORT (8000)
 CMD ["mcp.yaml"]
 
 # =============================================================================