claude-statusline-burnrate

claude-statusline-burnrate

Stop opening /usage. It's all right there.

Real server-side rate_limits — the same numbers /usage shows — answers what you actually want: am I ok, and how hard can I push? Pure bash + jq. No node, no daemons, no estimates.

The pacing math counts awake hours, so the trend doesn't fall "behind" every night while you sleep. And the colors stay gray until something needs you: only a meter that has actually warmed up gets to pull your eye.

Install

Needs jq and Claude Code v2.1+.

Easiest — you already have an AI open. Paste this into Claude Code:

Using the Antigravity CLI too? Add "and the same statusLine in ~/.gemini/antigravity-cli/settings.json" to the prompt (see below).

By hand — two commands:

Then add to ~/.claude/settings.json:

Open a new session. That's it.

Try it first

Prefer one line?

v1 is still here as statusline-classic.sh: same math, one row, emoji labels and an animated cat whose mood follows the worst meter.

Upgrading from v1: re-running the install gives you the card. The card is about 65 columns wide, so on a narrow terminal the classic line fits better.

Antigravity CLI (agy)

Both scripts also work as the status line of Google's Antigravity CLI. Needs jq, python3 and agy.

Already installed for Claude Code? Skip the download and just add the setting below. Otherwise:

Then add to ~/.gemini/antigravity-cli/settings.json:

agy's payload carries no rate_limits, so the script falls back to agy -p "/usage": parsed with python3, refreshed in the background every 3 min, and cached in ~/.claude/.cache/agy-quota.cache. Gemini models use the Gemini quota, Claude/GPT models the other one. The effort comes from the model name (Gemini 3.8 Flash (High)), and Gemini Flash, Gemini Pro and GPT get their own colors (plus a mascot in…