mlco2/codecarbon

Health v2 · maintenance only

Adoption not part of the score

1905 stars · 316 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

CodeCarbon is a lightweight Python library and CLI that estimates and tracks CO2 emissions from local computing hardware (CPU, GPU, RAM) by measuring power consumption and applying regional carbon intensity. It helps quantify the environmental impact of running code, especially machine learning workloads, and recommends ways to reduce it.

Use cases

• track carbon emissions of my machine learning training runs
• measure energy consumption of python code
• estimate co2 footprint of gpu training
• monitor sustainability of compute workloads
• compare carbon efficiency of ml models
• track emissions from a command without changing code

When to choose

• you run ML training or heavy compute on local hardware and want to quantify its carbon footprint
• you want a simple pip-installable tracker with a Python API and CLI
• you need emissions reporting for sustainability or ai-ethics goals

When to avoid

• you need emissions from remote GenAI API calls (use EcoLogits instead)
• you need fine-grained measurement of disk, network, cooling, or display power
• you run exclusively on hardware where power metrics are unavailable

Facets

library · maturity active

monitoring analytics benchmarking cli machine-learning developer-tools data-science energy python windows cli carbon-emissions energy-consumption co2-tracking green-computing ai-ethics sustainability emissions-tracker linux macos

Member repositories

For agents

markdown · JSON · MCP: product_card(name="mlco2/codecarbon")