Skip to main content
Loom provides a full command-line interface via the loom command.

Init

Initialize the project — generates config and template files (run once):

Build

Process text into structured memory:

Chat

Chat with memory-augmented responses:

Interactive chat commands

When in interactive chat mode (loom chat):

Schemas

Inspect registered schemas:

Serve

Start the API server in the foreground:
When invoked without a subcommand, loom serve runs the server in the foreground (Ctrl+C to stop). For background service management, see the subcommands below.

Service management subcommands

Only one Loom daemon can run at a time. If a daemon is already running, start will refuse to launch a new one regardless of the port specified. Use stop first, then start with a new port. Port can be set via the --port flag (highest priority) or in configs/loom.yaml under server.port (default: 8666).
See Service management for details on daemon mode, auto-start, and log management.

Summary