Most people paste, hit send, and hope for the best.
Vague prompts burn tokens, rack up costs, and waste energy — all before the AI even responds. You wouldn't mail a letter without proofreading it. Why send a prompt without checking it?
Wasted spend
Filler words and rambling context inflate your token count. On GPT-4 class models, that's $15–$75 per million tokens — burning money on noise.
Hidden energy cost
Every token consumes electricity, emits CO₂, and uses water for cooling. A sloppy prompt can use 35× more energy on reasoning models.
Worse outputs
Ambiguous prompts produce ambiguous answers. The clearer you are up front, the fewer re-prompts you need — saving time, tokens, and frustration.
How it works
Three seconds. Zero AI calls.
Promptlee runs entirely in your browser. No servers, no API keys, no data ever leaving your machine.
1
Type your prompt
Write normally on ChatGPT, Claude, Gemini, Perplexity, Poe — or any site you add.
2
Click the P badge
A floating gold badge appears on every enabled site. One click (or Ctrl+Shift+P) triggers the analysis.
3
Get your score
See a 0–100 score, letter grade, line-by-line issues, token count, cost, energy, CO₂, and water impact — all in under a second.
See it in action
One click. Instant feedback. Zero AI calls.
Watch Promptlee analyze a prompt in real time — right inside ChatGPT, Claude, or any AI site you use.
Works on ChatGPT, Claude, Gemini, Perplexity & Poe. Add any custom site in Settings.
What you get
Everything you need to know, before you hit send.
Example analysis of a 340-token prompt on Claude Sonnet
82
Score
340
Tokens
$0.001
Est. Cost
0.34
Wh energy
0.16
g CO₂
0.61
mL water
Click any card to expand the full calculation breakdown.
Energy & Cost
See the Real Cost of Every Prompt
Estimated energy, carbon footprint, water usage, and dollar cost — using EPA eGRID data and real model pricing.
Before Promptlee
$35.64/mo
→
After Promptlee (23% fewer tokens)
$27.44/mo
⚡
4.75
Watt-hours
🍃
2.28
g CO₂
💧
8.6
mL Water
💰
$35.64
/ month (40/day)
Enterprise ROI
Calculate Your Team's Savings
Enter your team size and daily prompt volume. See what Promptlee saves you every month.
$205
Monthly Cost Savings
5.1M
Tokens Saved / Month
11.5 kg
CO₂ Reduced / Month
43 L
Water Saved / Month
22,000
Total Prompts / Month
$891
Current Monthly Spend
23%
Avg. token reduction with Promptlee
40%
Filler words in average first drafts
Features
Built for anyone who talks to AI.
Whether you're a developer, marketer, student, or just someone who chats with Claude every day.
Instant efficiency score
A 0–100 score with letter grade based on 20+ rule-based checks: filler words, redundancy, vagueness, missing structure, placeholders, and more.
Energy & cost tracking
See estimated cost, energy (Wh), CO₂ (g), and water (mL) for every prompt. Click to expand the full math per metric.
Cumulative stats & trends
Track your average score, total tokens, and environmental footprint over time. Export everything as CSV.
Auto-analyze on send
Get feedback automatically every time you hit the send button. No extra clicks needed.
Prompt templates
One-click starter templates for coding, analysis, writing, research, and brainstorming. Start strong every time.
Works on any site
Ships with ChatGPT, Claude, Gemini, Perplexity, and Poe. Add any custom URL in Settings.
Dark & light themes
Matches your workflow. Warm dark by default, clean light if you prefer. Sidebar stays out of your way.
Enterprise ready
IT admins can push managed settings via Chrome Enterprise — enforce model tiers, lock score thresholds, set defaults org-wide.
Privacy
Your prompts never leave your device.
100% client-side. Zero telemetry.
Promptlee uses rule-based analysis — no AI model calls, no API keys, no server round-trips. Your prompts are analyzed in your browser's memory and never stored anywhere you don't control. History lives in Chrome's local storage and you can export or delete it at any time.
No account requiredNo data sent to serversNo third-party SDKsOpen source
Why Promptlee
Other tools analyze with AI. Promptlee doesn't need to.
That's the point. Checking your prompt shouldn't burn more tokens.
Promptlee
AI-based analyzers
Manual review
Uses AI tokens to analyze
✗ No
✓ Yes
✗ No
Instant (<1s) feedback
✓
3–15s
Minutes
Works offline
✓
✗
✓
100% private
✓
✗
✓
Cost & energy estimates
✓
Rarely
✗
Cross-site tracking & trends
✓
✗
✗
Free forever
✓
Freemium
✓
“I used to paste 500-word prompts into Claude without thinking. Promptlee showed me I was 40% filler. Now I write tighter prompts and get better answers, every time.”
— Early beta user, engineering team lead
Get started
Write better prompts, starting now.
Free. Private. Installs in 10 seconds. No account. No credit card.