Please say hello to CarbonGuru.io โ current tagline “AI makes world-class carbon expertise automatic, and available to everyone.”
Every product needs a community. I’ve been “building out loud” across more than one professional space for just over a year, and I couldn’t have reached a first-release moment like this without the inspiration and insight shared by many of you. Your generosity is appreciated, and I hope my gratitude has been visible in the mix of anecdotes, distillations, and occasionally on-topic observations I’ve shared along the way. Apologies for the salty language in the agent conversation snippets โ it’s the working-with-LLMs equivalent of percussive maintenance.
Snapshot: CarbonGuru brings business-level capabilities for understanding climate impact and offsetting carbon โ through Frontier Climate and CNaught โ to everyone. Accurate, actionable information on the climate impact of most products, services, and business activities was unavailable before, or unusable in everyday settings. CarbonGuru uses AI to fill that gap with detailed analysis in seconds, and pairs it with vetted, affordable, immediate action you can link directly to life, work, and business. This turns complicated accounting for spending and offsetting into a simple climate impact readout that feels just managing like your bank balance.
An old friend who leads a climate design practice asked why now is the right time for this โ a great way to combine Bruce Sterling’s MAYA rubric (“Most Advanced Yet Acceptable”) with the more basic question of genuine human value into a lens that’s actually useful for product-making. My answer: globally, the overwhelming majority of people already understand we need effective, positive action on the environment, starting yesterday (seeย 89percent.org). CarbonGuru uses newly available technology to let essentially anyone take meaningful action โ regardless of budget or the systemic leverage they have โ that’s directly relevant to them, and benefits everyone else in the world at the same time.
There’s a lot of “new” in this first public version while (hopefully) still hewing to a clear and compelling vision of benefit for users. Three of the core mechanics are new to the domain, if not unique. First, the model itself: taking climate action buys back your capacity to measure more of it. Each plan tier resets your analysis quota early once your offset purchases cross a dollar threshold, so verified real-world action โ not just a subscription โ is the currency for platform access. Second, it’s built for AI agents, not only people: an MCP server lets an assistant like Claude or ChatGPT analyze a product, scan your receipts, and actually buy and retire a verified credit in a single conversation, with no checkout screen โ an early instance of real-money climate action as a native agent capability. Third, the methodology is a scale leap: CarbonGuru applies Environmentally-Extended Input-Output analysis โ long confined to academic research and corporate reporting โ to any individual purchase, no matter how small, and complements carbon footprints with social cost of carbon calculation. The outcome is CarbonGuru makes managing your climate impact as easy as managing your bank balance.
None of that was affordable to build a year ago. Machine sense-making and AI marked a genuine inflection in the unit economics โ applying the science, product development, IT, and operations are now something like three orders of magnitude cheaper. Some of the core approaches CarbonGuru uses to scale LCA were first showcased at NeurIPS climate-and-ML workshops and published just over a year ago. The analytics platform was built from scratch by five generations of frontier coding agents โ this is where the salty language comes in โ and the data pipeline scales to intake new sources anywhere on the Internet via a self-composing agent that’s half open-source models, half proprietary harness, and half Claude. Like many attempts to seed a new category, CarbonGuru is a classic case of “last year this was a research project; now it’s a first-run product.”
Once you have the technology, the question becomes psychology. What will people do?
So far: personal customers using the preview have set up offset subscriptions for regular household spending โ groceries, family travel, ongoing things like kids’ camps and education programs (video walkthrough). As a software startup, I use it to automatically track and offset the footprint of the AI tools and services that build and run CarbonGuru itself (walkthrough with Claude). And the first business-tier customers are running higher-volume analysis of their transactions and operations.
To keep up with the exponential side of that โ if there’s an S-curve here to ride โ CarbonGuru needs support. Friends and family has started, I’m open to angels now. I’d welcome conversations with anyone interested in helping it grow to its full potential.
From the community, I’d ask three things:
- Try it and share feedback โ your perspective is valued.
- Send it to friends and colleagues. Make a connection โ who should see CarbonGuru? For partnerships, or funding?
- Act directly: Join the first group of customers. Become an investor!
Thanks for your attention, and for taking action.