Skip to content

swm Documentation

One CLI. Ten GPU clouds. Search, provision, deploy, sync, and manage — all over SSH.

GPU Search

Find the cheapest GPU across 10 clouds in one command. Filter by type, VRAM, price, and count.

Browse GPUs →

Pod Management

Create, start, stop, and terminate GPU pods from your terminal. Works across every provider.

Pod commands →

Workspace Sync

Push from one cloud, pull to another. Three-tier smart sync with inotify watcher and tar mode.

Learn more →

Agent Integration

Drop in a SKILL.md and your AI agent manages GPUs for you. Cursor, Codex, Claude Code, Windsurf.

Set up agents →

Lifecycle Guard

Monitors SSH, GPU util, filesystem writes, and processes. Auto-saves and terminates idle pods.

Guard docs →

Cost Tracking

Per-session, per-GPU cost tracking. Reconcile against provider billing APIs. Set budget alerts.

Track costs →

Frameworks

One command to install vLLM, Open WebUI, ComfyUI, SwarmUI, Axolotl, and more. Auto-detects GPU count.

See frameworks →

CLI Reference

Full reference for every swm command — gpus, pod, setup, sync, costs, guard, models, storage.

All commands →