CN-code check against Annex I: iron and steel, aluminium, cement, fertilisers, hydrogen, electricity, precursors included. The 50-tonne annual de minimis is computed from your own import lines, not estimated: below it you owe nothing, and you can show why. Above it, the authorised-declarant question starts your timeline.
In scope, or shown why not.
Embedded emissions belong to the installation that made the good, and to the precursors it consumed. Your supplier answers through a link, in the structure the regulation expects: installation identity, production route, direct and indirect emissions, precursor quantities. The same request grammar as every Regtrue engine: one request, named gaps, nothing chased by email.
Ask the installation. Not the trader.
Every emissions figure lands with its source recorded: verified installation data, supplier-declared, or default values, and the record keeps them apart. Where actual data is missing, the calculation falls back to defaults and says so. Data, evidence and compliance stay three separate things here, like everywhere in Regtrue.
Actual, declared, or default. Never blended silently.
The calculation shows what share of your number stands on actual installation data and what share still stands on defaults. It does not round that up. A partial number, honestly partial, is worth more than a complete-looking one nobody can defend.
The number is partial. The record says exactly where.
The annual CBAM declaration and the certificate arithmetic are outputs of the same record: goods, installations, emissions, evidence, decisions. This is the part in active development. The truth layer under it (who made what, with which emissions, proven how) is the same infrastructure already carrying PPWR and EUDR in production.
Built in the open. The check fills as the engine does.
Dates rendered from one verified source with citations.