GreenCalculus is an API for sourced emission factors and audit-traced carbon calculation.
Why GreenCalculus exists
Most online carbon calculators used by sustainability professionals fail one or more of three basic tests: they cite no emission factor source, they apply outdated Global Warming Potential values (typically AR4 or AR5 long after AR6 superseded them), and they produce a final number that cannot be traced back through a documented formula. For a quick consumer footprint quiz, none of that matters. For a corporate GHG inventory or a regulatory disclosure, all three are disqualifying.
GreenCalculus exists to close that gap. Every tool on the platform is built to be cited in a report — which means every output has to survive a methodology review by someone who reads the GHG Protocol for a living. The platform’s job is to make standards-aligned calculation infrastructure cheap enough to use casually and rigorous enough to use in formal disclosures.
It began as calculators and is now primarily infrastructure, because the requirement turned out to be the same one. Cite the factor, date it, show the formula: that is what a working practitioner needs from a calculator, and it is exactly what a developer needs from an emission-factor API. The teams shipping carbon accounting, ESG and CSRD reporting, procurement and spend analysis, and travel and logistics software were all rebuilding that layer themselves — and then maintaining it, every year, as publishers revised their numbers. The API exists so they do not have to. The developer surface is where the platform is going.
GreenCalculus is calculation infrastructure, not a consultancy. It does not produce certified audit outputs, professional environmental advice, or legal compliance opinions. Outputs — from the API and from the on-site calculators alike — are estimates derived from published emission factors. Users operating under formal disclosure regimes — CSRD, SEC, SBTi, national ETS — are responsible for confirming applicability with a qualified advisor.
The API is the product
A request to api.greencalculus.com returns an emission factor together with the exact place in the publisher’s own document it was read from, the publishing body, the licence that governs its reuse, its published uncertainty, and the data version it was served at. A request to one of the calculation endpoints returns the emissions figure and the working — the factor applied, the gas, the GWP set, the formula, and a receipt that lets the same result be reproduced later, with or without us.
“The exact place” is deliberately not “the cell”. DEFRA publishes a spreadsheet, so a DEFRA row cites a spreadsheet cell. The IPCC publishes chapter tables, CBAM an Annex with CN codes, Environment and Climate Change Canada a model database. Each row cites whichever of those its publisher actually issued. Search below and you will meet more than one of those shapes — that variety is the corpus being honest about its sources, not a gap in it.
Browsing is keyless and unmetered, deliberately. The factors can be read and checked without an account, because a data product whose numbers cannot be inspected before purchase is asking for trust it has not earned. Every tier can run every calculation engine, including PCAF — what the paid tiers buy is scale, redistribution rights, and the right to drop the attribution line, never access to the maths. That is a written commitment, not a launch promise.
Try it — this box is the API
Not a screenshot and not a sample response. Your browser calls
api.greencalculus.com directly and prints what comes back,
with no key and no account.
Try
No key, no account, no signup — and nothing on it is pre-rendered: if the API is down, the box says so rather than showing you a number it cannot fetch. The four figures below are read the same way, at render time from the deployed data layer rather than typed into this page. Each links to the surface that breaks it down.
What we hold
How it gets called
Most people who need a sourced factor are in a spreadsheet
So the same corpus is a Google Workspace add-on, verified by Google and free to install. =GC_FACTOR("grid.gbr.electricity.location_based") returns the number; the next cell returns the publisher, the reference, the version and a link to the public proof page. When the publisher revises a factor, the cell revises with it.
Reading is keyless. Ranges past roughly fifty rows want an API key — Apps Script calls from IP addresses shared with every other add-on, and the free route throttles there. Free and keyless for a workbook; a key for a model.
A calculation returns all of this, plus the arithmetic
Reading a factor is the easy half. What decides whether a number survives review is what happens when it is used — so a calculation response carries the row it applied and everything attached to that row, not just a total. Below is one such row, read from the deployed data layer at the moment this page rendered. A call to /v1/calculate/electricity returns exactly these fields alongside your kilowatt-hours, the multiplication, and a receipt that reproduces it later:
| Field | Live value |
|---|---|
| MasterBrain key | grid.gbr.electricity.location_based |
| Name | UK grid electricity — location-based (generation) |
| Value | 0.13096 kg CO2e per kWh |
| GWP basis | AR5_100 |
| Gas | CO2e |
| GHG Protocol scope | scope2 |
| Source | UK Government GHG Conversion Factors 2026 |
| Published by | Department for Energy Security and Net Zero (DESNZ) |
| Source cell | 'UK electricity'!E25 |
| Retrieved | 2026-06-18 |
| Licence | Open Government Licence v3.0 |
Read live from MasterBrain v2026.203. Verify against the publisher: UK Government GHG Conversion Factors 2026.
Multiply that by your kilowatt-hours and every term in the sentence is already cited. The calculation endpoints do exactly this and return the receipt with the answer — they need a key, on any tier including the free one, because a metered call is the only part of the API we count.
You just ran one endpoint. There are six more.
Factor lookup, plain-language resolve, activity emissions, embodied carbon, freight, business travel and PCAF financed emissions — the playground runs all of them against live data, and the free tier covers every engine including PCAF. No card.
Open the playground →Start with the free tier
Unlimited keyless browsing, every calculation engine including PCAF, and 1,000 metered calls a month. No card, commercial use with attribution.
Open the developer surface →The standards we calculate against
GreenCalculus does not author standards — it implements them. Every calculator derives its formula logic, emission factors, and GWP values from named, versioned, publicly auditable global standards. The platform groups them into three layers: the frameworks that define what to count, the scientific basis that determines what to multiply by, and the disclosure regimes that determine who reads the result.
What to count
What to multiply by
Who reads the result
A further set of standards — including ISO 14067 (Product Carbon Footprint), ISO 14040/14044 (LCA), the F-gas Regulation, the EU ETS, and the Australian Safeguard Mechanism — are referenced where calculator scope requires them. Full standards index →
How we publish a calculator: the four rules
The platform applies the same four publication rules to every calculator, every methodology page, every data reference and every API response. They are enforced at the development level — a page that fails any of the four is not published, and an endpoint that fails any of the four does not ship.
No black-box calculations
Every formula appears on the calculator’s methodology tab in full. Inputs map to outputs through explicit, published equations. No hidden multipliers, no proprietary adjustment factors, no “calibrations” applied silently to a final number. The API holds to the same rule structurally rather than by documentation: working is a field on every calculation response, so the derivation travels with the number instead of living on a page the caller has to go and find.
Reproducible by anyone with the standard
Any user with the same inputs and the same cited emission-factor version should be able to reproduce any GreenCalculus output independently. Worked examples — inputs, applicable factor, step-by-step derivation — accompany every methodology page. See for instance the natural gas combustion derivation or the Scope 2 market-based worked example. Over the API this becomes mechanical: each result carries a deterministic receipt and the data version it ran against, so it can be re-derived years later without depending on us still being here.
Uncertainty stated, boundary stated
Where estimation is necessary — activity data gaps, regional proxy factors, cross-sector boundary assumptions — uncertainty ranges are stated explicitly. Every tool documents which sector, geography, and activity type it covers and, just as importantly, where it should not be applied. Each API row carries the same information as data: a quality block with its representativeness and temporal lag, and its published uncertainty where the publisher gives one. What we do not hold is published as well, rather than answered with a silent proxy — see the absence record.
Versioned, with deprecated factors retained
Emission-factor tables are reviewed annually at minimum, triggered by new EPA, DEFRA, and IEA publication cycles. Version control is maintained per tool. A public changelog documents every factor update. Deprecated emission-factor versions are retained for reference — not silently replaced — so historical calculations remain traceable. Every API response names the data version that served it, and any version the archive still holds can be asked for again by name, so an upstream revision never moves a figure that has already been filed.
What we cover (and what we don’t)
Coverage is organised by the GHG Protocol scope framework, and each card below names the API engine that computes it. Two engines sit outside the scope frame, because the questions they answer are not scope questions: embodied carbon under EN 15978, and PCAF financed emissions for lenders and investors. Defining what the platform does not cover is part of the brief — the strongest sign a calculation tool is for general use is a long list of disclaimed applications.
Direct emissions
Stationary combustion, mobile combustion, industrial process emissions, and fugitive emissions from owned or controlled sources. Covers manufacturing, transport fleets, the built environment, and agricultural operations.
Purchased energy
Location-based and market-based methods for purchased electricity, steam, heat, and cooling. Regional grid emission intensities from IEA 2026. Full treatment of energy attribute certificates, RECs, REGOs, and power purchase agreements.
Value-chain emissions
All 15 GHG Protocol Scope 3 categories — upstream and downstream. Data quality tiers noted per category. Includes purchased goods and services (Cat. 1), business travel (Cat. 6), use of sold products (Cat. 11), and end-of-life treatment (Cat. 12).
Land & carbon removals
Soil carbon sequestration, forestry and land-use change, biogenic carbon accounting, and engineered carbon dioxide removal (CDR). Fully aligned to the GHG Protocol Land Sector and Removals Standard (2026).
Net-zero targets & pathways
SBTi-aligned near-term and long-term target setting using the absolute contraction approach. Validation against Corporate Net-Zero Standard criteria. FLAG sector targets for land-intensive companies.
Breadth is the thing a country count does not measure
The tile above says how many countries the corpus touches. That number is true and it is the wrong question, so here is the one underneath it — read live from the same producer that answers /v1/coverage, not written here:
Coverage is uneven, and the country count hides it. 151 countries hold at least one of the 13 families a corporate inventory normally needs — but only 30 hold four or more, and two countries hold every family. A country in the list is not a country you can finish an inventory in.
Every country, scored family by family →
Also served as data at /v1/coverage.
How we verify what we publish
The four rules above describe what gets published. This section describes how the platform proves they have been followed. Four mechanisms run on every published page and every API response — they are not aspirational, they are part of the publication pipeline.
Source verification
Every emission-factor workbook is verified cell-by-cell against the published standard. If a DEFRA or EPA source workbook updates, the hash changes and the affected calculators are flagged for review before they can update their displayed factor. This rules out silent substitution.
Read the verification process →Calculation verification
Every calculator has a corresponding methodology page that shows the formula, the worked example, and the inputs/output pair used to test the calculator on each release. If the calculator and the methodology page diverge, the calculator does not deploy.
Browse methodology pages →Reviewer accountability
Every page on GreenCalculus.com names the human who authored it and the system that verified it. Corrections are logged publicly with the date, the page affected, the old value, and the new value. Anyone can audit the platform’s history of being wrong.
Read the public changelog →Result-level reproducibility
The three mechanisms above verify what we publish. The fourth verifies what a caller received. Every calculation the API returns carries a deterministic receipt hash alongside the factor value and the data version it ran against, so a saved result reproduces identically for as long as it is kept — with or without us. A result that was not saved can be re-resolved against the version it used. And when a factor is later revised or withdrawn, we say what changed and whether the earlier figure still holds; it is never silently restated underneath a number someone has already filed.
How reproducibility works →Who runs GreenCalculus
GreenCalculus was built out of a specific, repeated frustration with the state of online carbon calculation tools available to sustainability professionals: emission factors with no citation, GWP values lagging the current IPCC assessment by a full report cycle, and outputs that could not be traced back through any documented formula. For internal analysis or for a citation in a published disclosure, those tools were unusable.
Jeremiah personally signs off every methodology page before publication, maintains the MasterBrain data layer that supplies emission-factor values across the platform, and is the named author on every calculation. He also builds and operates the API itself — the gateway, the calculation engines, the MCP server and the client SDKs — so the person answering an integration question is the person who wrote the endpoint. The verification pipeline (GreenCalculus Engineering) audits the work; Jeremiah is responsible for it.
Cited by Wikipedia
Wikipedia editors apply a documented reliable-source policy (WP:RS) when admitting external citations. The two entries below are instances where GreenCalculus standards pages were selected as the published reference supporting a specific factual claim on the English Wikipedia. Each citation is checkable against the article’s reference list and revision history.
ISO 14064
Footnote 3Wikipedia’s ISO 14064 article cites GreenCalculus as the published reference for the 2026 regulatory landscape around third-party assurance — specifically the section describing how mandatory disclosure regimes including the EU CSRD and frameworks in Singapore, Australia, and Hong Kong require GHG-statement assurance under ISO 14064-3 or its IAASB equivalent ISAE 3410, with most regimes starting at limited assurance and trajectoried toward reasonable assurance by 2028–2030.
Kigali Amendment to the Montreal Protocol
Footnote 19Wikipedia’s Kigali Amendment article cites GreenCalculus as the published reference for the HFC phase-down baseline mechanism for non-Article-5 (developed) countries — specifically that the 15%-of-baseline target by 2036 is calculated as the average HFC consumption between 2011–2013, plus a buffer of 15% of the party’s HCFC baseline to account for the ongoing transition from ozone-depleting substances.
Where we publish our work
Three ways in, over one corpus. The REST API is the product; the MCP server is the same corpus addressed by an agent instead of a program; the SDKs are thin typed clients over the REST surface, not a separate capability. Nothing is behind a different dataset — a figure fetched by an agent and a figure fetched by a script are the same row, at the same version, with the same source reference attached.
The product. Sourced factors and the calculation engines over them; keyless and unmetered to browse.
Point any MCP client at it and your agent gets twelve tools, so it cites a factor rather than inventing one.
Official typed clients: pip install greencalculus, npm install greencalculus.
Behind those three surfaces the site itself is structured into nine content layers, each with a defined editorial purpose and schema contract. This is the reference half of the business, and it is not a marketing brochure for the API — it is the same data under public review. Calculators answer specific questions. Methodology pages prove the calculators’ work. Standards pages cite the external authority. Data pages back up the factors, and every published factor gets a citable page of its own. Every layer is browsable on its own.
Single-question tools. One input set, one tCO₂e answer, one cited factor.
How the platform applies each standard. Formulas, worked examples, boundary statements.
Reference pages on external standards. What they say, who issues them, what GreenCalculus pages rely on them.
Emission-factor dataset references. Per-fuel, per-mode, per-region lookup tables.
Multi-source workflow tools — readiness checklists, scenario comparisons, lookup utilities.
How-to walkthroughs that link calculators into reporting workflows. Written for sustainability officers.
GHG accounting terms defined to GHG Protocol convention. A–Z reference.
One page per published emission factor — value, unit, source, licence, and citation.
Side-by-side comparisons of standards, methods, and accounting choices.
Questions people ask before they trust a number
Reading the corpus is free and needs no account: the browse endpoint, every factor page, every calculator, and the Google Sheets add-on all work without a key. The free API tier adds 1,000 metered calls a month, no card, commercial use with an attribution line. Paid tiers buy scale, redistribution rights, and the right to drop that line — never access to the maths. Every tier can run every engine, including PCAF. That is written down.
From the bodies that publish them — DEFRA/DESNZ, the IPCC, the US EPA, Ember, Eurostat, ÖKOBAUDAT, GLEC, the EU Official Journal for CBAM, and others. GreenCalculus authors no factors. Each row records who published it, the exact place in their document it was read from, the date it was retrieved, the licence, and the data version it was served at. Every publisher and its licence is listed at /masterbrain/sources/.
Because publishers do not all publish spreadsheets. Sampled across nine key prefixes at data version 2026.192, 5.8% of rows cite a spreadsheet cell — those are mostly DEFRA. The rest cite an IPCC chapter table, a CBAM Annex clause with its CN code, an ECCC model database process, or the publisher’s own page. Rows carrying no source reference at all: none. The wording used to promise the narrowest of those four; it now promises what we actually keep, which is the harder promise.
Commercial use is allowed on every tier, free included, with attribution. Redistribution is a different question and the answer is set by the publisher, not by us or by your price: some licence their factors to us on terms that do not let us pass them on. Those rows say so, and the redistribution feed excludes them rather than leaving the licence review to you. The per-category answer is at /developers/trust/.
The data layer takes a new version and the old value stays reachable. Pin any call with ?as_of=<version> and you get the figure as it stood, so a filed number stays reproducible after the source moves. Every release is listed with what changed and what it affects in the data release log, and Factor Watch emails you when a factor you track moves.
That is the design. Every factor has a public, keyless verification page pinned to a version, showing the publisher, the source reference, the licence, and whether the figure has changed since — and the response hands you its URL, so the proof travels with the number rather than living on our marketing pages. The search box in §02 above runs against the live API from your own browser; nothing on it is pre-rendered.
Published, at /masterbrain/absence/ — the gaps and why each one is a gap, rather than a silent proxy dressed up as an answer. Coverage is also uneven by design: 151 countries have grid data, far fewer have all thirteen families a full corporate inventory needs. Coverage by country shows exactly which.
Jeremiah Say, founder and lead systems architect, is the named author on every methodology page and answers technical mail directly — there is no sales team in between. The review pipeline, the 30-day Tier-1 update rule, the corrections policy and the independence and revenue disclosures are all at /governance/. Accepted corrections are logged publicly and contributors are credited by name.
Evaluating the API for production?
Integration questions, coverage gaps and design-partner places go straight to the founder — there is no sales team in between.
Talk to the founder →Limits, errors, and how to report them
What outputs are. Estimates derived from published emission factors using documented calculation methodologies. Not certified audit outputs. Not professional environmental, legal, or financial advice. Accuracy is bounded by the quality of user-supplied inputs and the currency of the underlying factors.
Regulatory use. Users operating under specific disclosure frameworks — including but not limited to EU CSRD, SEC climate disclosure rules, IFRS S2, SBTi, or national emissions trading schemes — are responsible for confirming that the applicable emission-factor version and calculation boundary meet their regulatory requirements. The platform’s job is to produce defensible internal numbers; signing them off for a filing is the user’s responsibility, working with qualified advisors.
Using the API in a product. What each tier may do with the data — commercial use, redistribution as data, and dropping the attribution line — is set out per category at /developers/trust/, because the answer differs by publisher and not by price. Some publishers licence their factors to us on terms that do not let us pass them on; those rows say so, and the redistribution feed excludes them rather than leaving the licence review to you. Uptime is published at /developers/status/ and what we log is at /developers/security/.
If you find an error. Email jeremiah@greencalculus.com directly — API, integration and design-partner enquiries are welcome at the same address, and /developers/ points there deliberately. Include the page URL or the endpoint and request, the inputs used, the output received, and the standard or dataset you believe conflicts with the result. Methodology corrections receive a response within 3–5 business days, are logged publicly in the changelog, and contributors are credited by name. General enquiries, PR requests, and sponsored content proposals are not accepted at that address — use /contact/.
Terms. Full terms at /terms/; privacy at /privacy/; verification process at /governance/; API rights and continuity at /developers/trust/.