01
Write
Prompt like you normally do in Claude Code, Cursor, ChatGPT, or Claude. TerraWatch sits quietly in the corner.
Hitting your Claude Code or Cursor limits again? TerraWatch compresses your prompts so you get more done before the cap.
See the CO₂, water, and energy behind every message, then shrink it in one click. Works in ChatGPT, Claude, Gemini, Perplexity, and your terminal.
Add to Chrome · it's freeIt lives where you already type. No new tab, no new habit, just lighter prompts and a running tally of what they cost the planet.
Prompt like you normally do in Claude Code, Cursor, ChatGPT, or Claude. TerraWatch sits quietly in the corner.
Hit one button. Your prompt gets rewritten to use fewer tokens while keeping the exact same meaning.
See the energy, CO₂, and water cost of every prompt. Watch your savings compound over time.
TerraWatch sits next to the box you already type in. One click strips the filler and keeps every requirement, so the AI gets the same ask for less.
Try it in your browsercan you please help me understand what an API is and how it works
Real examples from TerraWatch's deep rewrite. Token counts are estimates.
Every prompt has a price. Most of it is paid in tokens nobody needed to send.
That exceeds the annual emissions of entire countries like Norway or Portugal. And it is accelerating.
Assumes a $150 monthly AI spend with 30–50% of tokens doing nothing. Multiply that by a team and it is a hire.
Complex reasoning can hit 33 Wh. That is a lightbulb left on for two hours.
Data centers drink water to stay cool. A lot of it.
More prompts before you hit your daily cap. Lower API bills. Lighter footprint. Same meaning, every time.
Add the extension in one click. Every prompt you send after that gets lighter.
Everything people ask before installing. If yours is missing, the privacy page and pricing page go deeper.
It rewrites your AI prompts to use fewer tokens while keeping the exact same intent, and shows the CO₂, water, and energy behind every message. Fewer tokens means more prompts before you hit a daily cap, lower API bills, and a lighter footprint.
The browser extension runs inside ChatGPT, Claude, Gemini, and Perplexity. In the terminal, the twc CLI connects to the same account so your Claude Code and Cursor workflows get the same rewrites and tracking.
No. Quick clean strips filler and redundant phrasing locally. Deep AI rewrites remove everything that does not carry meaning and keep every requirement. Pro adds a before and after diff so you can see exactly what changed, and the original is always one click away.
Free includes quick clean, prompt quality scoring, highlight preview, ten deep AI rewrites a day, expanding vague prompts into specs, and the lifetime savings dashboard. Pro is $8 a month and adds unlimited deep rewrites, a historical prompt log, the before and after diff view, a monthly impact report, and priority support. Cancel anytime.
They are estimates. We count the tokens you send and apply published per-token energy figures for large models, then convert energy into CO₂ and cooling water. Treat them as a consistent yardstick for comparing prompts, not a utility bill.
Your prompts stay on your device. Quick clean runs entirely in your browser. When you explicitly click deep rewrite, the prompt is sent to our server, forwarded to Anthropic for rewriting, and returned without being stored. If you're signed in, your savings stats (token counts, CO₂, water, never the prompt text) sync to your account to power your dashboard.
Quick clean works as soon as you add the extension from the Chrome Web Store. A free account unlocks deep rewrites and your savings dashboard, links the extension to the twc CLI, and is what you upgrade to Pro.