Skip to content

cli.md

Latest commit

 

History

History
244 lines (171 loc) · 7.04 KB

cli.md

File metadata and controls

244 lines (171 loc) · 7.04 KB

CLI Reference

This document describes the CLI surfaces for fleet-rlm.

Entrypoints

There are two command entrypoints:

  • fleet-rlm: Typer-based command group with subcommands for server modes, optimization, and terminal chat
  • fleet: Lightweight launcher for terminal chat and Web UI startup

fleet-rlm Commands

The fleet-rlm command provides the primary CLI interface.

Usage: fleet-rlm [OPTIONS] COMMAND [ARGS]...

Run fleet-rlm demos and experimental runtimes.

Commands:
  serve-api   Run the FastAPI server surface (used by `fleet web`).
  optimize    Run DSPy optimization workflows.
  chat        Start standalone in-process interactive terminal chat.
  daytona-smoke  Run a native Daytona smoke validation without invoking an LM.
  daytona-snapshot  Create or refresh the reusable Daytona base snapshot.

fleet-rlm optimize

Run GEPA offline optimization for a registered DSPy module.

Usage: fleet-rlm optimize [OPTIONS] MODULE [DATASET]

Arguments:
  MODULE    Registered module slug to optimize (use 'list' to see available modules). [required]
  DATASET   Path to JSON or JSONL dataset. [optional]

Options:
  --output-path, -o PATH   Where to save the optimized DSPy module artifact.
  --train-ratio FLOAT      Training split ratio for GEPA compilation. [default: 0.8]
  --auto TEXT              GEPA optimization intensity: light, medium, or heavy. [default: light]
  --report                 Print a markdown report summary after optimization.
  --help, -h               Show this message and exit.

Examples:

# Optimize a module with default settings
uv run fleet-rlm optimize my-module dataset.jsonl

# Run heavy optimization with a custom output path
uv run fleet-rlm optimize my-module dataset.jsonl --auto heavy --output-path ./artifacts/optimized.pkl

# Generate an optimization report
uv run fleet-rlm optimize my-module dataset.jsonl --report

fleet-rlm serve-api

Run the FastAPI HTTP/WebSocket server. This is the backend for the Web UI and is invoked by fleet web.

Usage: fleet-rlm serve-api [OPTIONS]

Options:
  --host TEXT      Bind host [default: 127.0.0.1]
  --port INTEGER   Bind port [default: 8000]
  --help           Show this message and exit.

Examples:

# Start server on default host:port (127.0.0.1:8000)
uv run fleet-rlm serve-api

# Bind to all interfaces on custom port
uv run fleet-rlm serve-api --host 0.0.0.0 --port 8080

fleet-rlm chat

Start standalone in-process interactive terminal chat with the RLM agent.

Usage: fleet-rlm chat [OPTIONS]

Options:
  --docs-path PATH         Optional document path to preload as active context
  --trace / --no-trace     Enable verbose thought/status display
  --trace-mode TEXT        Trace display mode: compact, verbose, or off
  --volume-name TEXT       Optional Daytona volume name for persistent storage
  --help                   Show this message and exit.

Trace Modes:

  • compact: Condensed trace output
  • verbose: Detailed thought/status display
  • off: Disable trace output

Examples:

# Start interactive chat
uv run fleet-rlm chat

# Preload a document as context
uv run fleet-rlm chat --docs-path ./docs/architecture.md

# Enable verbose trace output
uv run fleet-rlm chat --trace --trace-mode verbose

# Persist chat state to a specific Daytona volume
uv run fleet-rlm chat --volume-name my-volume

fleet-rlm daytona-smoke

Run a native Daytona smoke validation without invoking an LM.

Usage: fleet-rlm daytona-smoke [OPTIONS]

Options:
  --repo TEXT   Repository URL to clone into the Daytona sandbox.  [required]
  --ref TEXT    Optional branch or commit SHA to checkout after clone.
  --help        Show this message and exit.

Examples:

# Validate the experimental Daytona setup against a repository
uv run fleet-rlm daytona-smoke --repo https://github.com/qredence/fleet-rlm.git --ref main

fleet-rlm daytona-snapshot

Create or refresh the reusable Daytona base snapshot used by fleet-rlm.

The default snapshot name is fleet-rlm-base. Normal sandbox creation prefers this snapshot for faster startup and falls back to a declarative image build when the snapshot is absent or not active.

Usage: fleet-rlm daytona-snapshot [OPTIONS]

Options:
  --name TEXT        Name of the reusable Daytona base snapshot to bootstrap.
                    [default: fleet-rlm-base]
  --base-image TEXT  Base OCI image used when creating or refreshing the
                    snapshot. [default: python:3.12-slim]
  --refresh          Delete an existing snapshot with this name/id before
                    recreating it.
  --help             Show this message and exit.

Examples:

# Ensure the default reusable Daytona base snapshot exists
uv run fleet-rlm daytona-snapshot

# Rebuild the default base snapshot from the configured default image
uv run fleet-rlm daytona-snapshot --refresh

# Bootstrap a custom snapshot name
uv run fleet-rlm daytona-snapshot --name fleet-rlm-base-experimental

fleet Launcher

The fleet command is a lightweight launcher that provides quick access to terminal chat and the Web UI.

usage: fleet [-h] [--docs-path DOCS_PATH] [--trace-mode {compact,verbose,off}]
             [--volume-name VOLUME_NAME]
             [{web}]

Start standalone fleet interactive chat. Hydra overrides are supported as
key=value tokens. Use 'fleet web' to launch the Web UI server.

positional arguments:
  {web}                 Optional subcommand (e.g., 'web' to launch the Web UI).

options:
  -h, --help            show this help message and exit
  --docs-path DOCS_PATH
                        Optional document path to preload into the chat session.
  --trace-mode {compact,verbose,off}
                        Trace display mode.
  --volume-name VOLUME_NAME
                        Daytona volume name for persistent storage.

fleet (Terminal Chat)

When run without arguments, fleet starts standalone terminal chat.

# Start terminal chat
uv run fleet

# Start with document context
uv run fleet --docs-path ./README.md

# Start with verbose tracing
uv run fleet --trace-mode verbose

# Start with a custom Daytona volume
uv run fleet --volume-name my-volume

fleet web

The web subcommand launches the Web UI server.

# Start Web UI on default port (127.0.0.1:8000)
uv run fleet web

The Web UI will be available at http://127.0.0.1:8000. The launcher delegates to fleet-rlm serve-api --host 0.0.0.0 --port 8000 while preserving Hydra overrides.

Hydra Overrides

Both fleet and fleet-rlm serve-api support Hydra-style configuration overrides as trailing key=value tokens. This allows runtime configuration changes:

# Example: override runtime settings
uv run fleet web dspy_lm_model=gpt-4

# Example: override runtime settings
uv run fleet volume_name=my-custom-volume

See Also