“Why is it repeating itself?”
Your working context is filling up.
Context telemetry for AI
See context burn in real time.
Know when your model is about to forget.
One-time purchase Chrome extension Runs locally
Context remaining~31,200 tokens left
⌁ NORMALContext window128Ktokens
Tokens used96,800tokens
Est. room31,200tokens
01 / THE INVISIBLE PROBLEM
One minute your model knows the brief. The next, it drops constraints, repeats itself, or forgets a decision from 40 messages ago.
TokenMeter makes the invisible visible—before the quality falls off.
Your working context is filling up.
Earlier instructions may be dropping out.
Now you have a signal, not a hunch.
02 / ZERO-FRICTION TELEMETRY
Context remaining~31,200 tokens left
⌁ NORMAL03 / NO FAKE PRECISION
AI chat websites do not expose every hidden system prompt, tool payload, truncation rule, or provider tokenizer. TokenMeter estimates the visible conversation locally and shows the result as a practical fuel gauge—not as billing data.
Ship with context awareness
A tiny instrument for people doing serious work with AI.
Secure checkout by Stripe · Promotion codes accepted · Download unlocks after verified payment
04 / FAQ
No. The estimate is calculated in your browser from visible page text. The extension contains no analytics or external network requests.
No browser overlay can see hidden provider context. TokenMeter is deliberately an estimate and clearly labels itself that way.
The first release supports ChatGPT, Claude, Gemini, and Perplexity. Its selectors are designed with fallbacks because these sites evolve.
After payment, download and unzip TokenMeter. Open Chrome Extensions, enable Developer mode, choose Load unpacked, and select the folder.