shelly mcp configure

shelly mcp configure

Configure AI assistant MCP integration

Synopsis

Configure AI assistant MCP integration.

This command writes the necessary configuration to enable AI assistants to use the Shelly CLI via MCP (Model Context Protocol).

Supported AI assistants: –claude-desktop Configure Claude Desktop app –claude-code Configure Claude Code (VS Code extension / CLI) –gemini Configure Gemini CLI

You can specify multiple flags to configure all at once.

1
shelly mcp configure [flags]

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
  # Configure Claude Desktop
  shelly mcp configure --claude-desktop

  # Configure Claude Code
  shelly mcp configure --claude-code

  # Configure all supported assistants
  shelly mcp configure --claude-desktop --claude-code --gemini

  # Preview changes without writing (dry run)
  shelly mcp configure --claude-desktop --dry-run

  # Use custom shelly binary path
  shelly mcp configure --claude-code --shelly-path /usr/local/bin/shelly

Options

1
2
3
4
5
6
      --claude-code          Configure Claude Code
      --claude-desktop       Configure Claude Desktop
  -n, --dry-run              Preview changes without writing files
      --gemini               Configure Gemini CLI
  -h, --help                 help for configure
      --shelly-path string   Path to shelly binary (auto-detected if not specified)

Options inherited from parent commands

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
      --config string           Config file (default $HOME/.config/shelly/config.yaml)
      --log-categories string   Filter logs by category (comma-separated: network,api,device,config,auth,plugin)
      --log-json                Output logs in JSON format
      --no-color                Disable colored output
      --no-headers              Hide table headers in output
      --offline                 Only read from cache, error on cache miss
  -o, --output string           Output format (table, json, yaml, template) (default "table")
      --plain                   Disable borders and colors (machine-readable output)
  -q, --quiet                   Suppress non-essential output
      --refresh                 Bypass cache and fetch fresh data from device
      --template string         Go template string for output (use with -o template)
  -v, --verbose count           Increase verbosity (-v=info, -vv=debug, -vvv=trace)

SEE ALSO

  • shelly mcp - MCP server for AI assistant integration