🍃 Your Climate Dashboard
Understanding your AI usage footprint — not to judge, but to inform.
Conversations
23
across 5 days
→ typical
Total tokens
48.2k
input + output
↑ 35% vs last week
Estimated energy
36 Wh
≈ 0.036 kWh
↑ 28% vs last week
Estimated CO₂
24g
EU average grid
↑ 28% vs last week
Daily usage
⚠️ Thursday was 3.5x your daily average — you had a long code generation session (8 conversations, ~18k tokens).
What does 24g CO₂ look like?
≈ 1.6 kettles
of boiling water
of boiling water
Boiling a kettle
1.6x
Toasting bread
0.8 slices
Hot shower
16 seconds
Emails sent
6 emails
Driving
200 meters
Usage by model
llama3.1:70b
62%
llama3.1:8b
24%
mistral:7b
14%
💡 62% of your usage was on the 70B model. For simpler questions, the 8B model uses ~7x less energy.
Suggestions based on your usage
You often request long-form output
4 of your conversations this week generated 1000+ word responses. If bullet points work, they use ~60% less tokens. Try asking: "Give me the key points" instead of "Explain in detail."
Repeated similar questions on Thursday
You asked 3 variations of the same coding question. Refining one conversation is more efficient than starting fresh each time.
Good use of the smaller model
You used llama3.1:8b for quick lookups this week — that's a great fit. It used 7x less energy than the 70B model for similar results.
Choose your equivalence
Pick the comparison that makes most sense to you. We'll use it throughout the dashboard and in session tickers.