This document describes the current CLI surfaces for fleet-rlm.
There are two command entrypoints:
fleet-rlm: Typer-based command groupfleet: lightweight launcher for terminal chat and Web UI startup
uv run fleet-rlm --helpCurrent command set:
init: install scaffold assets to a Claude directoryserve-api: run FastAPI serverserve-mcp: run FastMCP serverchat: run standalone in-process terminal chat
uv run fleet-rlm init --helpKey options:
--target PATH(default~/.claude)--force--skills-only--agents-only--teams-only--hooks-only--no-teams--no-hooks--list
uv run fleet-rlm serve-api --helpOptions:
--host(default127.0.0.1)--port(default8000)
Hydra-style config overrides are supported as trailing key=value tokens.
uv run fleet-rlm serve-mcp --helpOptions:
--transport(stdio,sse,streamable-http; defaultstdio)--host(default127.0.0.1)--port(default8001)
uv run fleet-rlm chat --helpOptions:
--docs-path PATH--trace / --no-trace--trace-mode TEXT(compact,verbose,off)
uv run fleet --helpBehavior:
fleetstarts standalone terminal chat.fleet webstarts the Web UI/API server on0.0.0.0:8000viafleet-rlm serve-api.
Options:
--docs-path PATH--trace-mode {compact,verbose,off}--volume-name TEXT--secret-name TEXT
If other docs reference commands not shown in current --help, treat them as historical.
For MCP setup examples, see Using the MCP Server.