Transparency Report
Our AI footprint,
openly shared.
Moves uses AI to extract event details from social media posts. This page shows our complete environmental impact — app pipeline and development costs — updated monthly.
Loading...
What That Looks Like
📱
—
iPhone charges
💡
—
hours of a 9W LED bulb
🚗
—
miles driven equivalent
🌳
—
tree-days to absorb
Breakdown — App vs. Development
✦ App Pipeline
Model
—
Input tokens
—
Output tokens
—
Est. API calls
—
CO₂e
—
Energy
—
⌨ Development Sessions
Model
—
Input tokens
—
Output tokens
—
Est. sessions
—
CO₂e
—
Energy
—
Offset Commitment
⏳ Pre-Revenue
We're not offsetting — yet.
Efficiency Practices
Monthly Log
| Month | App CO₂e (g) | Dev CO₂e (g) | Total CO₂e (g) | App Tokens In | Notes |
|---|
Why we publish this.
Most apps that use AI don't talk about what it costs — energetically or environmentally. We think that silence is a problem. Moves exists to help people live more fully in their city. Doing that while quietly ignoring our own footprint would be a contradiction we're not willing to make.
We include development costs alongside app costs because transparency means showing the full picture. Building Moves cost more than running it this month — and that's something you deserve to know.
This page updates every month. If the numbers go up, you'll see it here.
Sources & methodology. CO₂ and energy per Haiku query: 0.22 Wh and 0.10g CO₂e (CarbonCredits.com, March 2026). Sonnet estimated at 3–5× Haiku per token based on model size ratio. Water: 0.32 ml per prompt (Toolpod.dev, 2025). iPhone charge: 15 Wh. LED bulb: 9W. Car: 404g CO₂/mile (US EPA average). Tree absorption: ~21 kg CO₂/year per tree. App API call count estimated by dividing input tokens by average prompt size (~750 tokens). Dev session count estimated from Anthropic console model breakdown. All figures are estimates — Anthropic does not publish per-token carbon figures publicly.