Open-source tool that tracks and predicts energy use and carbon footprint of model training.
- Pricing
- Free
- Category
- AI Energy Monitoring
- Compliance
- Not stated
Verified 2026-10-22Website
AI Energy Monitoring
Tracarbon
Python library and CLI that tracks device energy consumption and estimates carbon emissions.
Verified 2026-10-22
Overview
Tracarbon is a Python library and CLI that tracks a device's energy consumption and estimates its carbon emissions. It detects the device and location automatically, can be run from the command line to collect and export measurements, or used through a Python API, and ships exporters for Datadog, Prometheus and Kubernetes. Carbon intensity is fetched from the Electricity Maps API when a key is configured, otherwise bundled data for supported countries and cloud regions is used. It supports per-process tracking, Prometheus metrics for Kubernetes containers, macOS energy readings via IOReport, Linux RAPL power readings for Intel and AMD, and NVIDIA GPU power on Windows, and is released under the Apache 2.0 licence.
Use cases
- Track a device's energy use and emissions
- Export carbon metrics to Prometheus or Datadog
- Measure per-container emissions in Kubernetes
- Attribute emissions to individual processes
Sources
More in AI Energy Monitoring
Category page →Open-source Python library that estimates the carbon emissions of compute.
- Pricing
- Free
- Category
- AI Energy Monitoring
- Compliance
- Not stated
Verified 2026-10-22Website
Python library for tracking the CO2 emissions of machine learning models.
- Pricing
- Free
- Category
- AI Energy Monitoring
- Compliance
- Not stated
Verified 2026-10-22Website