Workflows

Workflows

Task-oriented recipes for tu, derived from tu --help. Every command and flag below matches what the CLI ships. New to tu? Start with the install guide for setup, including multi-machine sync.

Command grammar

tu reads a positional command line of the form:

tu [source] [period] [display]

• Source — which tool’s usage to report:
  • cc — Claude Code
  • codex / co — Codex
  • oc — OpenCode
  • all — every source (default)
• Period — the time window:
  • d / daily — daily (default)
  • m / monthly — monthly
• Display — how to present it:
  • (bare) — snapshot
  • h / history — history
• Combined — period and display fused into one token:
  • dh — daily history
  • mh — monthly history

Any token may be omitted; omitted tokens fall back to their defaults (all, daily, snapshot).

Common recipes

tu                   # Today's cost, all tools (snapshot)
tu cc                # Today's cost, Claude Code
tu h                 # Daily cost history, all tools (pivot)
tu cc mh             # Monthly cost history, Claude Code
tu m                 # This month's cost, all tools

Output formats

By default tu renders a formatted table. For scripting and piping, data commands accept structured output formats (data commands only):

• --json — output data as JSON
• --csv — output data as CSV
• --md — output data as Markdown

tu --json            # today's cost, all tools, as JSON
tu cc mh --csv       # Claude Code monthly history, as CSV

Watch mode

Watch mode turns any data command into a live, continuously-refreshing display (data commands only):

• --watch / -w — persistent polling mode with live display
• --interval / -i <s> — poll interval in seconds (default: 10, range: 5-3600)
• --no-rain — disable the matrix rain animation in watch mode

tu -w                # live snapshot, refreshing every 10s
tu -w -i 30          # live, polling every 30 seconds
tu -w --no-rain      # live, without the matrix rain animation

Multi-machine

When multi mode is configured (see the install guide), these flags control cross-machine aggregation:

• --sync — sync metrics before fetching (multi mode)
• --user / -u <user> — show usage for a specific user (multi mode only)
• --by-machine — show per-machine cost breakdown (data commands only)

tu --sync            # pull/push metrics, then show today's cost
tu -u alice          # show usage recorded under the 'alice' profile
tu --by-machine      # break today's cost down per machine

Cache control

By default tu reads from cached data. To bypass the cache and fetch fresh data (data commands only):

• --fresh / -f — bypass cache, fetch fresh data

tu -f                # ignore cache, re-fetch today's cost