Skip to content

cli.md

Latest commit

 

History

History
94 lines (69 loc) · 2.71 KB

cli.md

File metadata and controls

94 lines (69 loc) · 2.71 KB

CLI Reference

The rookery CLI communicates with the rookeryd daemon over HTTP.

When api_key is set in ~/.config/rookery/config.toml, the CLI automatically attaches it to daemon requests. rookery auth generate prints a strong suggested key for the config file.

Global Options

--daemon <url>    Daemon address (default: http://127.0.0.1:3000)
--json            Output as JSON (supported on most commands)

Commands

Server

rookery start [profile]       # start server (default profile if omitted)
rookery stop                  # stop server
rookery sleep                 # unload the model but remember the active profile
rookery wake                  # wake the sleeping profile
rookery swap <profile>        # hot-swap to a different profile
rookery status                # show server state, PID, uptime
rookery bench                 # run PP + gen speed benchmark
rookery profiles              # list available profiles

rookery status reports sleeping when auto-sleep or manual sleep has unloaded the backend. The next chat request also wakes the server automatically.

GPU

rookery gpu                   # VRAM, temp, utilization, power, processes

Agents

rookery agent start <name>    # start an agent
rookery agent stop <name>     # stop an agent
rookery agent update <name>   # stop, update, restart
rookery agent status          # list agents with status
rookery agent describe <name> # detailed health, uptime, restarts, errors

Models

rookery models search <query>     # search HuggingFace for GGUF repos
rookery models quants <repo>      # list quants with sizes and VRAM fit
rookery models recommend <repo>   # VRAM-aware best quant recommendation
rookery models list               # locally cached models
rookery models pull <repo> [--quant Q4_K_M]  # download a model
rookery models hardware           # show hardware profile

Releases

rookery releases              # upstream release status (llama.cpp, vLLM)
rookery releases --json       # JSON output for scripting

Logs

rookery logs                  # last 50 log lines
rookery logs -n 100           # last 100 lines
rookery logs -f               # follow mode (stream via SSE)

Config

rookery config                # validate config, show resolved commands

Other

rookery auth generate        # print a strong rky-... API key
rookery completions <shell>   # generate shell completions (bash, zsh, fish)

JSON Output

Most commands support --json for machine-readable output:

rookery status --json | jq '.state'
rookery agent status --json | jq '.agents[0].uptime_secs'
rookery gpu --json | jq '.gpus[0].vram_used_mb'