Green Module¶
Accountability-first, always-on carbon footprint and environmental impact tracking for AI inference.
The Problem¶
Every LLM inference request has an environmental cost:
- Carbon emissions from GPU power consumption
- Water usage for data center cooling
- Embodied carbon in hardware manufacturing
This cost is largely invisible. Providers don't disclose per-request emissions. Users have no way to understand, track, or reduce their AI environmental footprint.
OpenClaw surfaces this by default — not as guilt, but as accountability and awareness.
Key Features¶
| Feature | Description |
|---|---|
| Always-on tracking | Captures carbon data from the first request |
| Per-request granularity | Atomic traces that aggregate into summaries |
| Conservative estimates | Worst-case factors backed by academic research |
| Confidence scoring | Every estimate carries a 0.0–1.0 confidence |
| Standards compliance | GHG Protocol, CDP, TCFD, ISO 14064, SBTi |
| Multiple interfaces | CLI, Gateway dashboard, REST API |
Quick Start¶
# Check your environmental impact
openclaw green status
# View intensity metrics (TCFD)
openclaw green intensity
# Export for GHG Protocol reporting
openclaw green export --format ghg-protocol --period 2025-Q1
Architecture¶
Next Steps¶
- Installation — Enable and configure tracking
- Quick Start — 5-minute walkthrough
- Core Concepts — Understand traces, factors, confidence
- CLI Reference — All commands documented
- Standards — Compliance reporting guides