Skip to content

Docker

The container is the sandbox: no network, and it sees only the folder you mount.

docker pull ghcr.io/kd5ryn/b123d-print-mcp

Then point your client at Docker instead of uvx (Claude Desktop shown):

{
  "mcpServers": {
    "b123d-print": {
      "command": "docker",
      "args": ["run", "-i", "--rm", "--network", "none",
               "-v", "/Users/you/parts:/work", "ghcr.io/kd5ryn/b123d-print-mcp"]
    }
  }
}
  • -i is required (MCP talks over stdin/stdout); --rm removes the container afterwards.
  • Your parts and exports appear in the mounted folder (/Users/you/parts above). Settings (printer, hole compensation) are kept there too, in .config/b123d-print/, since the container has no lasting home folder. The image sets B123D_PRINT_CONFIG_IN_WORKSPACE=1 for this (see Settings).
  • On Linux, add "--user", "1000:1000" (your id -u:id -g) so exported files belong to you.
  • The image turns the OS sandbox off (B123D_PRINT_SANDBOX=none): bubblewrap cannot run inside a container without extra privileges, and the container already isolates the scripts.
  • Tags: latest and each release (ghcr.io/kd5ryn/b123d-print-mcp:0.2.0); pin one if you want updates only when you choose.

Building the image yourself

For unreleased code, or to check what goes into the image:

git clone https://github.com/KD5RYN/b123d-print && cd b123d-print

docker build -t b123d-print-mcp .

Then use b123d-print-mcp instead of ghcr.io/kd5ryn/b123d-print-mcp in the client config.

To run the image as a network service for several clients, see Over the network (HTTP). To update, see Updating.