Command Line Interface¶
dcclib-cli ships a single entry point (cli) with five top-level command groups:
dcclib [--version] [--help]
├── validate – Validate a DCC XML file (XSD + Schematron)
├── convert – Convert between DCC XML and JSON
├── extract – Extract formulas or embedded files from a DCC
├── mcp – Start the MCP server for AI agent integration
├── transform – Apply XSLT stylesheets
└── signature – Sign and verify XML digital signatures
Installation¶
# Recommended: install the CLI globally with pipx
pipx install "dcclib[cli]"
# with MCP support for AI agents
pipx install "dcclib[cli,mcp]"
# and with unit-aware quantities support using dccQuantities (requires .NET on Windows and libmono on Linux)
pipx install "dcclib[cli,quantities]"
# or everything for the full experience
pipx install "dcclib[cli,mcp,quantities]"
After installation, the dcclib command is available in your PATH:
dcclib --help
dcclib --version
Model Context Protocol (MCP)¶
dcclib-cli provides Model Context Protocol (MCP) support via pycli-mcp, allowing AI agents to interact with the CLI as structured tools.
Every subcommand is exposed as a tool (e.g. validate_xsd, convert_to_json), derived from its options and arguments.
# serve over stdin/stdout (default, for clients that launch the server themselves)
dcclib mcp
# serve streamable HTTP at http://127.0.0.1:8000/mcp
dcclib mcp --transport http --host 127.0.0.1 --port 8000
For step-by-step setup instructions in LM Studio, see the LM Studio (MCP) Guide.
Common options¶
Output format (--format / -f)¶
Most commands support a --format (or -f) option to control the output style:
| Format | Description |
|---|---|
plain |
Human-readable text (default) |
table |
ASCII table layout |
json |
Machine-readable JSON |
junit |
JUnit XML (for CI integration) |
Commands¶
The full, auto-generated command reference (every subcommand, argument, and option) lives on the Command Reference page: