★ GitHub

Know when your AI gets dumb.

A live meter for Claude Code that shows how full your context window is, so you /compact before your agent forgets your code.

~/my-appmainSmart12%💀
👆 click the meter. make your AI dumber.
Install free →★ GitHub
01

Watch it climb.

One session, top to bottom. You see the drop coming before the answers go bad.

~/my-app ⎇ mainSmart
12%fresh context. agent is sharp.
~/my-app ⎇ mainCoasting
45%still solid. keep going.
~/my-app ⎇ mainCooked
78%starting to drift. plan your /compact.
~/my-app ⎇ mainDumb 💀
96%re-inventing APIs. /compact now.
~/my-app ⎇ mainSmart
8%↑ after /compact, fresh.
02

Context rot is real.

Early on, your agent one-shots the function. Three hundred messages later it re-introduces the bug you killed an hour ago and invents an API that does not exist. You cannot feel it coming. The meter can.

// the fix is /compact. you just need to know when.
03

Five states. One glance.

Empty window = sharp. Full window = mush. The bar names exactly where you are.

Smart
0-24%
Coasting
25-49%
Foggy
50-69%
Cooked
70-89%
Dumb
90-100%
04

Three things. That is it.

// reads your % locally

The real context-window percentage, straight from Claude Code, on your machine. No guessing, no network call.

// never crashes

Every failure path falls back to safe output and exits clean. 64 tests prove it.

// one file, zero deps

MIT, no account, no telemetry, no config. Install it and it works immediately.

ZERO TOKENS

It reads a number Claude Code already computed. It never calls the model. It will never cost you a cent or a single token of context.

05

Questions, answered.

What is dumbometer?
A Claude Code status-line gauge that shows how full your context window is, from Smart to Dumb, so you /compact before quality drops.
What is context rot?
As a session fills the context window, the model spreads its attention thinner and starts making mistakes it would not make fresh. dumbometer measures the cause, context percent, so you see it coming.
Does it cost tokens?
No. It reads a number Claude Code already computed and runs locally on every render. Zero tokens, zero network, zero cost.
Will it slow down or break Claude Code?
No. The hot path is synchronous and sub-millisecond, and every failure path falls back to safe output and exits clean. 64 tests cover the never-crash contract.
Does it work with the 1M context window?
Yes. It reads the real window size, 200k or 1M, and shows the percentage either way.
How do I install it?
Three commands: /plugin marketplace add MaximoCorrea1/dumbometer, /plugin install dumbometer, /dumbometer:setup. It is free and MIT.
06

30 seconds. Then forget it exists.

Paste three commands into Claude Code. The meter shows up.

1/plugin marketplace add MaximoCorrea1/dumbometer
2/plugin install dumbometer
3/dumbometer:setup
07

Three tiers. Pick the free one.

The product is free and that is the honest answer. The other two exist because we could not help ourselves.

👈 the one you want
Free
The whole product.
$0 /forever
  • The meter, live in your status line.
  • All 5 states, from Smart to Dumb.
  • Reads your real context-window %.
  • Zero tokens. Runs locally. Never crashes.
  • MIT. No account, no telemetry.
Install free →
this is everything. you want this one.
Big Dumb Energy one-time
A tip jar that does stuff.
$9 once
  • 🎨 Extra color themes: vaporwave, matrix, hazard.
  • 🔥 Roast Mode. "/compact, champ."
  • 🥚 Rare galaxy brain 🌌 easter-egg frames.
  • 🏆 dumb --score. flex your peak dumb score.
  • 🔤 Custom labels. rename "Dumb" to anything.
Get it. $9 once →
a tip jar with perks. you do not need it. but it is nine bucks.
Singularity
Not a real tier. Probably.
$1M /yr
  • We fly to your home office.
  • We tap your shoulder when the context fills.
  • Weighted blanket included.
Contact sales 😉
we will not be doing this. please buy the free one.