IT Carbon Emission Monitoring & Reduction
We help you measure the carbon footprint of your software stack, including AI workloads, and act on it, using established open-source tools rather than a closed black box.
Our Approach
What We Actually Do
We don't ask you to trust a single proprietary score
Assess Your Software Stack
We measure the carbon footprint of your applications and infrastructure, including AI training and inference workloads, which are often the largest and least visible source of emissions in a modern stack.
Integrate Established Open-Source Tools
Rather than building a closed measurement engine from scratch, we integrate tools the Green Software community already trusts: the Green Software Foundation's Impact Framework, CodeCarbon, and the Green Metrics Tool, through our own open-source project, Carbonara.
Score It Against a Recognized Standard
Carbonara has a built-in assessment based on the Software Carbon Intensity (SCI) specification, so your result is comparable against a published, vendor-neutral methodology, not an in-house metric only we understand.
Help You Move to Lower-Carbon Hosting
Where hosting region or provider is a meaningful lever, we help you identify and move workloads to less energy-intensive regions, based on your actual measured footprint, not a generic recommendation.
Built On, Not Instead Of
The Open-Source Tools We Work With
Established community tools we integrate, plus Carbonara, which we build and maintain
Our own open-source project: a developer tool that brings carbon assessment into the code editor and CI pipeline, integrating the tools below so you get feedback earlier in the development lifecycle.
The Green Software Foundation's open-source framework for modeling software environmental impacts. We use it as a foundation rather than reinventing impact modeling.
An established open-source Python package for tracking the carbon emissions of running code, widely used in the ML and data science community.
An open-source tool for measuring the energy and carbon impact of software, which we work with alongside the tools above depending on what your stack needs.
Built-In SCI Assessment
The Software Carbon Intensity (SCI) specification, published by the Green Software Foundation, defines a standard, vendor-neutral way to calculate a rate of carbon emissions for a software system.
Carbonara includes an SCI assessment built on this specification, so your score reflects a published methodology rather than an internal formula. Because the SCI specification leaves real choices to the implementer (which software boundary to draw, which functional unit to scale by, and whether to use measured or modeled data), a generic default configuration rarely fits every organization well.
We help you make those choices deliberately: defining the right software boundary and functional unit for your systems, and deciding where measured data is available versus where a modeled estimate is the honest option, so the resulting score is one you can defend and act on rather than one that just looks good.
How We Work With You
From first measurement to an ongoing, defensible number
-
Assess Your Stack
We measure your current software and AI workload footprint using the tools above, rather than starting from assumptions.
-
Define Your SCI Boundary
We help you choose the software boundary and functional unit that make your SCI score meaningful for your actual systems.
-
Identify the Real Levers
Hosting region, workload scheduling, and architecture choices affect emissions very differently depending on your stack. We prioritize based on your measured data, not a generic checklist.
-
Monitor Over Time
Carbonara integrates into your editor and CI pipeline, so carbon impact becomes a visible, ongoing signal instead of a one-off report.
How This Relates to OSA
Two different levels of carbon analysis that some organizations need both of
This service measures the carbon footprint of your software and infrastructure directly, at the level of code, deployments, and compute. Our OpenSustainability Analysis Framework (OSA) works at a different level: it analyzes what companies, including your own suppliers or portfolio holdings, have already disclosed about their carbon emissions in sustainability reports.
If you need to understand your own software's footprint, this page is the relevant service. If you need to analyze what other companies report about their emissions, OSA is the relevant tool. Some clients use both together, for example, to compare a supplier's disclosed emissions against an independent technical assessment of systems you can directly measure.
Related Reading
View all posts »More on carbon tracking, software sustainability, and the Carbonara project.
Research & Development Collaboration in Climate Science
Work with leading experts for innovative research collaborations in climate science.
ESG Consulting & Sustainability Roadmaps
A pragmatic sustainability roadmap: build the in-house data you need, fill gaps with curated sources, and use that intelligence to drive environmental, social, and governance decisions, including your business model.
ESG & AI: Revolutionizing Real Estate Investment
Discover how AI is shaping the future of ESG in real estate investments for a sustainable and profitable tomorrow.
Climate Risk Assessment Services
A pragmatic path from materiality analysis to a data foundation you control to custom vulnerability models for the assets that matter most.
Frequently Asked Questions
Did you build Impact Framework, CodeCarbon, or Green Metrics Tool?
No. These are established open-source projects maintained by their own communities (the Green Software Foundation, in the case of Impact Framework). We integrate them through Carbonara and help you apply them correctly, rather than claiming to have built them ourselves.
What is Carbonara, exactly?
Carbonara is our own open-source project: a developer tool that brings carbon assessment into code editors and CI pipelines, integrating the tools above so feedback arrives earlier in the development lifecycle than a post-deployment audit would.
What is the SCI, and why does it matter?
Software Carbon Intensity is a specification published by the Green Software Foundation for calculating a rate of carbon emissions for a software system. It matters because it's a published, vendor-neutral methodology, rather than a number only one company can reproduce.
Can this cover our AI training and inference workloads specifically?
Yes. AI workloads are often the largest and least visible source of software-related emissions in a modern stack, and the tools we integrate are widely used in the ML and data science community for exactly this.
How is this different from your OpenSustainability Analysis Framework (OSA)?
This measures your own software and infrastructure directly. OSA analyzes what companies report in their sustainability reports. They answer different questions and are complementary, not competing.
Can you help us actually reduce emissions, not just measure them?
Measurement is the starting point. Where hosting region, scheduling, or architecture changes would meaningfully reduce your measured footprint, we help you prioritize and implement those changes.
Tell Us About Your Stack
We'll tell you honestly what we can measure today, and what would need custom work.