Carbon Sight uses public-domain methodology only. No proprietary or commercial LCA databases are referenced. Every number traces to a citeable source.

Read the full public methodology → Public audit & commitments →

The three-tier system

Tier 1 — Audit-grade (green). Manufacturer-published Product Environmental Reports, Environmental Product Declarations, or peer-reviewed LCAs. Each item carries a direct source URL. Examples: Apple iPhone 15 (61 kg, sourced from Apple PER), Starbucks Caffè Latte (0.55 kg, sourced from Starbucks sustainability disclosure). 239 items currently (2.3% of catalog).

Tier 2 — Material-emission-factor estimate (blue). Weight × material composition × emission factor. Material emission factors drawn from EPA USEEIO v1.3 and DOE GREET 4.0 — the same datasets used by the California Air Resources Board for LCFS pathway evaluation. 9,902 items (95.6% of catalog).

Tier 3 — Per-unit reference (amber). Bulk emission intensity per kilogram for raw commodities. Sourced from Poore & Nemecek 2018 (Science) and similar peer-reviewed literature. 214 items (2.1% of catalog).

Primary data sources

GREET 4.0 — Argonne National Laboratory, U.S. Department of Energy. Process-level life-cycle model for fuels, vehicles, and materials. argonne.gov/greet

USEEIO v1.3 — U.S. EPA Supply Chain GHG Emission Factors. 1,016 NAICS commodities at kg CO₂e per dollar producer-price output. epa.gov/land-research/us-environmentally-extended-input-output-useeio-models

USLCI — U.S. Life Cycle Inventory Database (NREL). nrel.gov/lci

Poore & Nemecek 2018 — "Reducing food's environmental impacts through producers and consumers." Science 360 (6392), 987-992. The benchmark dataset for food carbon intensity.

Manufacturer EPDs / PERs — Direct sources for Tier 1 items. Each carries its source URL.

Open Food Facts — Real-time API for branded grocery items. Queried at scan time. world.openfoodfacts.org

Catalog audit history

Audit v1.0 (May 17, 2026). Hand-verification of 15 high-visibility consumer reference items against published LCAs. Pre-audit accuracy: 73% within published ranges. Identified systematic methodology error in 200 prepared-food items where per-kg-of-ingredient emission factors were applied to total dish weight (over-estimating beverages, soups, prepared dishes by 3–5×). Patched 24 most consumer-visible items with literature-sourced corrections. Each patched item carries an audit_note field with reference citation. Post-audit accuracy on the same items: 100%.

What this is and isn't

This is a methodology infrastructure for product carbon footprint estimation using audit-traceable public-domain data. Suitable for: consumer education, internal sustainability tracking, preliminary modeling, scenario comparison, and the basis for further verification.

This isn't a substitute for third-party verification of formal product claims. Brands using Carbon Sight for externally-communicated certifications (e.g., shelf labels, marketing claims) should obtain independent verification under ISO 14067, GHG Protocol Product Standard, or equivalent accredited body. Carbon Sight provides the preliminary modeling and methodology infrastructure, not accredited verification.

Operator credentials

Chain Reaction Labs, Inc., the operator behind Carbon Sight, has submitted thousands of public filings involving life-cycle carbon intensity calculations to state and federal regulatory agencies. These submissions are subject to agency audit, technical challenge, and public comment — the same evidentiary standard applied to corporate GHG disclosures and product carbon claims. That is the level of LCA rigor we bring to Carbon Sight's methodology.

Numbers without sources are vibes. Sources without uncertainty are lies. Every estimate here carries a range, a tier, and a traceable methodology. Found something wrong? Email info@carbonsight.io with the reference.