Methodology

Mula explains what each figure measures, the information it uses, its sources and version, and the limits of the result. Missing information stays visible. Detailed implementation rules and data that cannot be redistributed are not published here. Figures are calculated without AI; brand-stated lifespan and uploaded LCA studies remain distinct from Mula calculations.

This page is available as a file. The file names the factor library version in force when it was taken, together with the date, so a copy already attached to a review file still says the same thing after the library version rises.

How each figure is calculated

For each figure: what it means, what it is not, then the detail for anyone who wants to check the arithmetic. Every formula runs in Mula's rules engine; no AI fills in or estimates any figure.

Recycled material

This figure: The share of the product's material that the brand declares as recycled, as a percentage and as a weight.
What it is not: Not a statement about waste that was actually diverted, and not zero for a line the brand has not filled in: a silent line is left out of the arithmetic and named in the coverage line.

Technical detail

Formula: Percentage = the sum over composition lines that declare a recycled content of (line share × line recycled content), divided by the sum of the declaring lines' shares. Kilograms = product weight × the sum over the same lines of (line share × line recycled content). Kilograms are not weight × percentage: the part of the composition that declares nothing is given no rate at all.
Inputs: Composition per line (fibre and share), the recycled content per line as the brand typed it, and the product weight. A fibre that is itself a recycled variant, such as recycled cotton, counts as 100% on its line when no figure is entered, because the brand declared it by choosing the fibre; an entered figure always wins.
When it is not computed: Not computed when no line declares a recycled content, or the product has no composition lines. The kilograms are empty when the product weight is empty. What is empty is never read as zero.
How it is shown: The engine rounds the percentage to 1 decimal place and the kilograms to 3. On the passport there are no decimals: a whole percentage, and a value above zero but below one per cent shows as less than one per cent, never as zero; a mass below one kilogram shows in grams. The passport card carries a coverage line, based on so many of so many materials, when a line is silent, and is withheld entirely when there is no basis.
Limitations: Recycled content is the brand's statement, with its evidence tier; Mula does not measure it. A zero the brand declares stays zero per cent, and that is a different state from a silent line.

Recycled material mass

This figure: The weight of recycled material in this product, in kilograms, as the brand declared it: one kilogram of recycled material counts as one kilogram, with no factor applied.
What it is not: Not waste avoided, not virgin material left unused, and not an impact prevented: a one-to-one ratio does not show that this recycled material displaced virgin material on the market. It is the same quantity as the kilograms of Recycled material, shown as a mass.

Technical detail

Formula: Recycled material mass = kilograms of recycled material, one to one.
Inputs: The same as Recycled material: composition, recycled content per line, product weight.
When it is not computed: Empty when the kilograms of recycled material are empty. The passport card is also withheld when the value is zero, because zero kilograms is not a fact that earns its own card.
How it is shown: Written with a ± sign on the passport, and a mass below one kilogram in grams. The ± marks the figure as an estimate from the brand's declaration; it is not a computed interval, and no uncertainty range is published for this metric.
Limitations: Until 25 September 2026 this figure was named Textile waste avoided, and that name was withdrawn from every surface because it stated an outcome that is not measured: recycling losses (cutting and shredding lose material) and the market substitution rate are unknown, so no factor is applied and nothing is declared avoided. The figure did not change; its name did.

Material carbon footprint

This figure: An estimate of the greenhouse gas emissions from making this product's materials, from field or refinery to the point where the material is ready to use, in kilograms of CO2e for one item.
What it is not: Not the carbon footprint of the whole product: spinning for some materials, weaving or knitting, dyeing, sewing, transport, use, and end of life are not included. Not a measurement of this brand's supply chain: the factors are published figures from outside Indonesia.

Technical detail

Formula: Kilograms of CO2e = the sum over composition lines of (product weight in kilograms × line share × the emission factor of that line's fibre, in kilograms of CO2e per kilogram). The factor is chosen from the fibre code through its default factor key. Recycled content never touches this formula: a recycled fibre has its own factor, so nothing is counted twice.
Inputs: The product weight as the brand typed it, the composition per line, and the factor library in force, the table below.
Boundary: From the cradle to the material's factory gate, per kilogram of material, hundred-year global warming potential. The gate follows each row's source and is not uniform: synthetics are counted to the pellet, before they become fibre; natural fibres to spinning-ready fibre; recycled cotton to cutting and shredding only; lyocell to fibre at the facility gate including its dissolving pulp; silk to reeled raw silk; leather to tanning. The point is written in each row's source column.
When it is not computed: Not computed when the product weight is empty, when there are no composition lines, or when even one material has no sourced factor. A partial sum is never shown, because it would report a smaller footprint than the truth. The public passport uses sourced factors only; the dashboard, reports, and API still render figures from provisional factors with their label and version.
How it is shown: The engine rounds to 2 decimal places. On the passport there are no decimals: a ± sign, the anchor materials only, cradle to gate, and a mass below one kilogram in grams of CO2e. On the dashboard and in reports, two decimals. The passport only publishes a figure over a composition that adds up to exactly one hundred per cent, because publishing requires it; on the dashboard an incomplete composition is computed over the declared shares only, so the figure runs low until the composition is complete.
Limitations: No uncertainty is computed; the ± sign marks an estimate, not an interval. The characterisation method and the treatment of biogenic carbon follow each row's source and are not uniform: the Idemat rows use IPCC 2021 GWP100 with no biogenic uptake, and rows from published studies state their own method in the source column, including those that carry a biogenic uptake or credit. Factors are point values for the geography written on their row, none of them Indonesian. The weight used is the finished product's: losses from fibre to fabric to item are not counted, so the footprint of the material actually bought to make it is larger than this figure.

Circular potential

This figure: A Mula indicator based on material composition, declared recycled content, care and repair information, and a documented product return route. Fibres are grouped by family. Evidence tiers label the supporting information without changing the calculation. A product with no composition gets no score; its data gap remains visible.
What it is not: Not a measurement of recycling achieved or a scientifically validated environmental score. The result follows versioned Mula product rules, not an LCA study.

Rules v2, in force since 23 September 2026.

Technical detail

Formula: The assessed dimensions are combined under versioned Mula rules. Implementation weights, thresholds and calibration are not published on this page.
Inputs: Composition and recycled content per line, care or repair information, components and their disposal routes, and the product return route. Missing inputs remain distinct from zero.
When it is not computed: Not calculated without composition lines; the passport card is withheld.
How it is shown: The passport shows Low, Medium, or High; the dashboard also shows points.
Limitations: A return-route entry does not establish that the route operates. Evidence tiers stay visible; a brand declaration does not become evidence of performance simply because it produces a band.

Processing carbon

This figure: Emissions from the processes that treat the fabric, such as printing or dyeing, calculated from the energy a workshop uses per kilogram of fabric multiplied by the emission factor of that energy. This figure shows on the dashboard, in reports, and in the API, not on the passport, while the energy factor library is still labelled provisional.
What it is not: Not part of the Material carbon footprint, and never silently added into it: where both are shown, both stand with their own labels.

Technical detail

Formula: Processing carbon per kilogram of fabric = the sum over energy carriers of (energy per kilogram of fabric × that carrier's emission factor). Product carbon = that figure × the product's fabric mass in kilograms. A carrier without an amount or without a factor becomes a named gap, never zero.
Inputs: The process type, energy per kilogram of fabric per carrier (electricity in kWh, diesel in litres, LPG and paraffin in kilograms), the fabric mass, and where the figure came from: a supplier statement, a bill or meter reading, or an estimate. Carrier factors come from Mula's energy factor library, the same one the Scope 1 and 2 module uses: grid electricity from ESDM DJK for the Java-Madura-Bali system, fuels from IPCC 2006. That grid factor is the only Indonesian data point anywhere in Mula's factor libraries.
Boundary: Gate to gate for the named process only, per kilogram of finished fabric. The fibre is not included, nor are the manufacture of dyes and auxiliaries, sewing, transport, use, and end of life.
When it is not computed: A named process with no energy figure at all shows as not yet measured. For screen and digital printing only there is a default from one published study on Finnish grid electricity, labelled as a default with its region; it is used only when the brand has entered no figure at all for that process, is replaced wholesale once measured energy is entered, and a default is never averaged with a measurement. No other process has a default.
How it is shown: Its own provenance line, processing, measured energy, with the factor's region and the energy library version. The factor values are written on the dashboard and in reports, beside each figure, with their source and version.
Limitations: The grid electricity factor in use was published in 2019 and is labelled provisional until its latest vintage is confirmed. Energy per kilogram allocated from a shared meter carries its allocation assumption, written by the brand on the row. The printing defaults stand on Finnish grid electricity, far lower in carbon than Java's, so they run low for an Indonesian workshop and are never rendered publicly.

Calculations use energy data and record the factor sources and versions alongside results. This public page does not distribute the numerical factor catalogue or internal mappings.

Brand-level figures

This figure: Brand figures on the dashboard, in reports, and in the API are simple averages per unit: the sum of the per-product figures divided by the number of products that could be computed, one unit of each product, always with a coverage line naming how many products it was computed from.
What it is not: Not a brand total. A total needs the number of units made per product, which is not recorded yet, and this average is weighted by nothing because the only meaningful weight is that number.

Technical detail

Formula: Average per unit = the sum of the per-product figures ÷ N, over the N products whose figure could be computed. Products that could not be computed are left out and named in the coverage line, computed from N of M products, never counted as zero. Products without a composition are left out of the Circular potential band counts and named in that coverage line.
Inputs: Per-product figures from the same engine as the passport, and the factor library version in force when computed.
When it is not computed: Empty when not a single product could be computed. An average with no factor version to name it is not rendered at all.
How it is shown: Rounded to the same precision as the per-product figures, with a form line, simple average per unit, and a coverage line beneath. A released report is frozen at the form and factor version of its release and does not move when a library or rule set rises.
Limitations: An average per unit gives a product made a thousand times the same weight as one made once. It is useful for comparing a catalogue with itself over time, not for stating a brand's impact.

Brand-stated lifespan

This figure: The lifespan in the Life chapter is the brand's statement of how long its product is typically kept, shown exactly as the brand wrote it, a single number, a range, or a number with a plus sign, labelled as the brand's estimate, with the brand's basis where the brand wrote one.
What it is not: Not Mula's calculation, not a default, and not a guarantee. Mula does not calculate a lifespan, does not reshape the figure, and shows no card when the brand has stated nothing. The rule set for this metric is empty, and that is a recorded state: no published table turns category and fibre into the life of one particular product.

Finishing declaration

This figure: The brand declares four things about its product's finishes and added materials: dye class, finishing treatments, hardware materials, and elastic type, from a fixed vocabulary. On the passport each axis shows as declared, and an axis not yet filled in shows as not declared, in the same size type. On the dashboard the brand sees its coverage: how many of the four axes are declared, and which are not, named.
What it is not: Not a safety statement, not a score, and not a substance check. Coverage is a count of declared axes, and it never appears on the passport, because four of four on a page about one item reads as a statement about the item. An empty axis means the brand has not declared it, not that there is nothing there. Choosing no added finish or no hardware is the brand's answer and counts as declared; empty does not. No word promising safety or the absence of a substance appears on any surface, and there is no irritant flag by fibre class.

Where it was made

This figure: How many of a product's 4 making stages name a place, a party, or both, as the brand declared them: where the fibre comes from, where the fabric was made, where it was dyed, and where it was sewn. On the passport it shows as the place and party per stage, with a coverage line reading so many of 4 steps recorded so far.
What it is not: Not a score, not a percentage, and not a statement that the supply chain is fully traced or was checked by Mula. It counts what the brand recorded and names what is not recorded yet; a country name is never read as a sign of quality, and an empty stage is named as such, never as a warning.

Technical detail

Formula: The number of stages named, out of 4. A stage is named when its record carries a place at country level or finer, or a supplier name, or a maker whose attestation is active and whose scope includes the name; any one suffices, and a stage named by all three still counts once. No weighting between stages and no fraction.
Inputs: Fibre or yarn is read from the material origin on each composition line. Cloth, colour and finishing, and stitching are read from the product's three country columns (country of weaving or knitting, of finishing, of assembly) and from the components in the cloth, colour, and stitching roles: origin region, supplier name, and maker. Fastenings and trims are not counted as a stage; making them a fifth stage is a founder decision still open.
Boundary: Country level or finer means a country, province, regency, city, or village. Continents, regions, blocs, and the word imported on its own do not name a place and do not count; nor do stand-ins for not known. The whole text is compared, not the words inside it: imported from India names a country and counts. Australia counts, because it is a country's name. No list of countries filters the input; only sayings broader than a country are refused, so that a real place name is never refused silently.
When it is not computed: Not computed, and not shown, when the product has no composition lines, no components, and all three country columns empty. A product that has any of them but names no stage shows zero of 4 with all four stages as named gaps, not as a score of zero.
How it is shown: Passport: in the Origin chapter, one line per stage with its place and party or the sentence not recorded yet, an evidence tier per stage, and a coverage line so many of 4 steps recorded so far, with the same bar indicator as origin by component. Dashboard: the same lines, and a stage not yet recorded links to the field that fills it. Never a percentage, never a coloured bar.
Limitations: The place and party are the brand's statements with their evidence tier; Mula does not check that the cloth was in fact woven where it says. Evidence tiers label and never cut: a stage named by a certificate and one named by the brand's own words each count as one. A maker whose attribution is withdrawn or not yet attested does not count on the passport, because the passport's reader cannot see them. Rules v1, in force since 25 September 2026.

Evidence tiers

Every material or process claim carries its evidence tier. The tier labels the figure and never cuts its value: a hand-woven step counts in full whether it is evidenced by a certificate or by the brand's own statement, and the reader weighs it. The absence of a certificate is not a shortfall: certificates cost money, and the farmers and makers upstream rarely hold them. What is documented is stated with confidence.

CertifiedA third-party certificate, with its validity tracked.
DocumentedBatch records, photographs, statements from farmers and makers, or an attestation by a cooperative, group, or trained facilitator recorded by the brand.
Self-declaredThe brand's own account, with no supporting document.

The exception is the readiness score on the brand's dashboard, which does weigh evidence, because it measures how well documented a passport is, not the product. That score never appears on a passport.

Emission factors

Library version v2.1-idemat-2026RevB1.

Source references, system boundaries, geography and methods are shown below. The numerical factor catalogue and internal mappings are not published.

Materials without an eligible factor can still be recorded on a passport. Composition remains visible, while a carbon figure that cannot be calculated stays a gap.

This library carries no water factor for any fibre yet, so the water footprint is not computed and not shown anywhere until a citable water factor exists.

Limitations of the material factors

The material carbon figure is a sum over rows whose sources stop at different points, use different climate methods, and treat biogenic carbon differently. The lists below are read from each row's columns in the library in force, not typed, so they name exactly what is in the table today.

Boundary per row: Cotton: Cotton fibre before spinning, trade mix, without global sea transport.
Organic cotton: Organic cotton fibre before spinning.
Polyester: PET pellets, before fibre extrusion and spinning.
Recycled polyester: Mechanically recycled rPET pellets, ex transport.
Viscose rayon: Viscose fibre production, before spinning.
Linen: Flax fibres ready for spinning into linen.
Nylon: Nylon pellets, before fibre extrusion and spinning.
Recycled nylon: Chemically upcycled polyamide, before fibre spinning.
Elastane: Elastane (polyurethane) pellets, before fibre extrusion.
Acrylic: Acrylic fibres, before spinning.
Jute: Jute fibre including degumming, market mix, before spinning.
Wool: Raw wool from the farm, including transport to Rotterdam, before spinning.
Leather: Bovine leather after chrome tanning.
Bamboo viscose: Viscose fibre production, before spinning; applied to bamboo viscose on the assumption of the same process.
Ramie: Ramie fibre after carding, before spinning.
Hemp: Hemp fibres (12% moisture content), before spinning.
Recycled cotton: Mechanical cotton recycling, cutting and shredding only; excludes collection, sorting, and carding.
Lyocell: Lyocell fibre at the facility gate including dissolving pulp, before spinning.
Silk: Reeled raw silk from cocoons, from mulberry cultivation to the factory gate.
Characterisation method: IPCC 2021 GWP100: Cotton, Organic cotton, Polyester, Recycled polyester, Viscose rayon, Linen, Nylon, Recycled nylon, Elastane, Acrylic, Jute, Wool, Leather, Bamboo viscose, Hemp.
ReCiPe midpoint: Ramie.
GWP; vintase IPCC tidak dinyatakan sumber yang dikutip / IPCC vintage not stated by the cited source: Recycled cotton.
IPCC 2013 GWP100: Lyocell.
IPCC 2007 GWP100 (GWP100 v1.02): Silk.
Biogenic carbon: No biogenic carbon uptake deducted; fossil and process footprint per the Idemat method: Cotton, Organic cotton, Polyester, Recycled polyester, Viscose rayon, Linen, Nylon, Recycled nylon, Elastane, Acrylic, Jute, Wool, Leather, Bamboo viscose, Hemp.
Includes biogenic carbon uptake at the cultivation stage; the total stays positive, so it is not exactly comparable with the Idemat rows: Ramie.
Biogenic carbon treatment not stated by the cited source: Recycled cotton.
No biogenic uptake deducted; all five stage contributions are positive: Lyocell.
Credits the carbon bound in the silk fibre (45%), under 2% of the figure; other mulberry uptake is treated as returned to the atmosphere: Silk.
Region: Trade mix of China, India, and the USA: Cotton.
India: Organic cotton.
Not stated by the source: Polyester, Viscose rayon, Linen, Nylon, Recycled nylon, Elastane, Acrylic, Bamboo viscose.
China: Recycled polyester, Ramie.
Market mix, countries not stated by the source: Jute.
Australia, transported to Rotterdam: Wool.
Hides from Argentina: Leather.
United States: Hemp.
Spain: Recycled cotton.
Türkiye, one plant in Gaziantep: Lyocell.
Southern India (Karnataka, Andhra Pradesh, Tamil Nadu): Silk.

The sum is therefore not a footprint at one consistent boundary but a footprint to the point each source makes available. Synthetics counted to the pellet leave out fibre extrusion and spinning, so the synthetic share of a blend runs low relative to the natural fibres beside it. The difference between GWP100 vintages is small for CO2 and larger for methane; rows that carry a biogenic uptake or credit are not exactly comparable with rows that do not.

Not one row measures an Indonesian supply chain. A region the source does not state is written as such, not guessed.

Where the factors come from. The library in force is published in the Netherlands and names the process geography of each row, a trade mix or a single country of origin; rows from published studies name the country of the study. None measures production in Indonesia. That means the material carbon figures here are screening-level estimates built on factors from outside Indonesia, not measurements of the brand's supply chain, and the passport card says materials only, cradle to gate precisely to limit that claim. We write it down because a figure whose source its reader does not know cannot be checked by anyone.

Where the material figures come from

A provisional library uses opening figures that carry no source to cite; a sourced library uses figures from a published dataset, and every factor carries its own source and process code. When the library changes, what changes is the basis of the calculation, not the product.

Different LCA libraries give different figures for the same material, and the spread can reach threefold. What we show is the figure from one library that can be cited, not the only figure that exists. The source and version are written on every block, so any figure here can be traced back to its dataset.

Libraries that have been in force

v2.1-idemat-2026RevB1 · published 24 September 2026 · in force nowIdemat 2026 RevB1 (Stichting Sustainability Impact Metrics, a TU Delft spin-off), IPCC 2021 GWP100 carbon footprint, cradle-to-gate, per kg of material. Each fibre factor carries its own Idemat process code. The eighteen v2.0 rows are carried unchanged. Four factors are absent from Idemat and come from published studies cited on those rows: recycled cotton, ramie, and from 24 September 2026 lyocell (Sürmelihindi & Balcı 2026, CC BY 4.0) and raw silk (Astudillo, Thalwitz & Vollrath 2014). Used under the 2026 EULA. Data from www.ecocostsvalue.com.
Material factors from Idemat 2026 (Stichting Sustainability Impact Metrics), data from www.ecocostsvalue.com
v2.0-idemat-2026RevB1 · published 4 September 2026Idemat 2026 RevB1 (Stichting Sustainability Impact Metrics, a TU Delft spin-off), IPCC 2021 GWP100 carbon footprint, cradle-to-gate, per kg of material. Each fibre factor carries its own Idemat process code. Two factors, recycled cotton and ramie, are absent from Idemat and come from published studies cited on those rows. Used under the 2026 EULA. Data from www.ecocostsvalue.com.
v1.0-provisional · published 13 August 2026The figures in this table are still provisional. Every factor will carry a citable source before public launch; until then the numbers here are used as early estimates and marked as such.

How figures are displayed

One figure lives at three precisions, and all three come from the same arithmetic.

The engine rounds percentages to 1 decimal place, kilograms to 3 decimal places, and carbon to 2 decimal places, then keeps the result with the factor library version that produced it.
The passport never shows decimals: mass and carbon are rounded to whole numbers, and anything below one kilogram changes unit to grams instead of losing its precision. Percentages are rounded to whole numbers, except a value above zero but below one per cent, which shows as less than one per cent and never as zero, because zero is a different statement about the item.
The dashboard and reports show two decimals for carbon and mass, because someone deciding on a material needs the differences that disappear in whole kilograms.
The ± sign on the passport marks the figure as an estimate from factors, not a measurement. It is not a computed interval: no uncertainty range is computed or published for any metric.
A card with no basis is withheld entirely on the passport, not shown empty or as zero; on the dashboard the gap is named together with what is holding it. Each metric block carries one provenance line naming the factor library version and, for a licensed library, its attribution.

Equivalences, Did you know?

This figure: Beneath a product's material carbon figure, the brand's dashboard adds one line of comparison: about how many full phone charges emit that much CO2e, computed from the energy per full charge and the same Indonesian grid factor that processing carbon uses. The real figure always shows first, and the comparison is a second line, not a replacement.
What it is not: Not a reduction, not an absorption, and not a verdict. The comparison is always with something that emits the same amount, never with something that absorbs it, because a sentence that sets an emission beside an absorption reads as if the emission had been settled. It is never compared with another product or brand, carries no colour, is never rendered when the figure is withheld, never appears on a brand figure with coverage gaps, and is not on the passport yet.

Equivalence table v1, in force since 25 September 2026.

Technical detail

Formula: Charges = the product's kilograms of CO2e ÷ (kWh per full charge × the grid electricity factor in kilograms of CO2e per kWh). Rounding: below ten to one significant figure; ten to one hundred to the nearest five; above one hundred to the nearest ten. Always prefixed about, never a decimal, and below one charge the line is not rendered.
Inputs: Energy per full charge 0.0194 kWh, the US EPA assumption (Greenhouse Gas Equivalencies Calculator, updated 4 August 2026): 28.446 Wh consumed in twenty-four hours less 22 hours of maintenance mode at 0.411 W. The grid electricity factor comes from Mula's energy factor library, the Java-Madura-Bali grid row (ESDM DJK), which is labelled provisional; it is read at render time, never stored, so a newer vintage moves every line at once. EPA uses a US grid factor for its own calculator; Mula does not.
When it is not computed: Not rendered when the material carbon figure is withheld, when the grid electricity factor cannot be read by the reader's seat, or when the result is below one charge.
How it is shown: One line beneath the material carbon figure on the dashboard product card, headed Did you know?, with a small line naming the kWh per charge, the grid factor, both sources, and the equivalence table version. Not on the aggregate strip, not in reports, not on the passport.
Limitations: The EPA battery assumption (28.446 Wh consumed in twenty-four hours) sits above many current phones, so each charge counts heavier and the count comes out lower, the conservative direction for a how-many reading. The grid factor is the Java-Madura-Bali publication, labelled provisional until its latest vintage is confirmed. The comparison adds no information to the figure; it helps a reader feel the grams, nothing more.

The electricity factor source and version are recorded with the comparison. The comparison illustrates consumption scale.

Sources

Every source that touches a figure on this page, with its licence terms as written on its row. Figures from research reports and papers are cited as a figure with its citation; no table, figure, or long quotation is copied from any source onto this page or onto any other Mula surface.

The material factor library in force. Idemat 2026 RevB1 (Stichting Sustainability Impact Metrics, a TU Delft spin-off), IPCC 2021 GWP100 carbon footprint, cradle-to-gate, per kg of material. Each fibre factor carries its own Idemat process code. The eighteen v2.0 rows are carried unchanged. Four factors are absent from Idemat and come from published studies cited on those rows: recycled cotton, ramie, and from 24 September 2026 lyocell (Sürmelihindi & Balcı 2026, CC BY 4.0) and raw silk (Astudillo, Thalwitz & Vollrath 2014). Used under the 2026 EULA. Data from www.ecocostsvalue.com.
The Ramie row. Dong, Xian & Yi 2018, Aerospace 5(3):81, DOI 10.3390/aerospace5030081, CC BY 4.0; ramie fibre after carding, functional unit 1,000 kg of fibre, cradle-to-gate, China; excludes spinning (yarn is 3.79 in the same paper); includes biogenic carbon uptake at the cultivation stage, so it is not treated identically to the Idemat rows
The Recycled cotton row. Esteve-Turrillas & de la Guardia 2017, Resources, Conservation and Recycling 116:107-115, as summarised in VTT-R-00941-22 Table 2; mechanical cotton recycling, cutting and shredding, Spain; excludes collection and sorting, and excludes carding
The Lyocell row. Sürmelihindi & Balcı 2026, Sustainability 18(13):6778, DOI 10.3390/su18136778, CC BY 4.0; industrial lyocell fibre at the facility gate, declared unit 1 kg of fibre, cradle-to-gate including dissolving pulp, IPCC 2013 GWP100 with no biogenic uptake deducted (all five stage contributions are positive), site data from one plant March to September 2023, Türkiye; excludes spinning
The Silk row. Astudillo, Thalwitz & Vollrath 2014, Journal of Cleaner Production 81:158-167, DOI 10.1016/j.jclepro.2014.06.007, Table 3 p. 162, copyright Elsevier, one figure cited; reeled raw silk at the factory gate, functional unit 1 kg, cradle-to-gate from mulberry cultivation, observed farm practice, southern India; the recommended-practice figure of 52.5 in the same table is not used; accounts for the carbon bound in the fibre
The VTT reports cited on this page (VTT-R-00941-22 and VTT-CR-04462-14) carry their portal terms: duplication for research and education only, never for sale. Only a figure with its citation is taken from them.
Energy factors for processing carbon and the Scope 1 and 2 module: ESDM DJK, Faktor Emisi GRK Sistem Ketenagalistrikan, Java-Madura-Bali system, for grid electricity; IPCC 2006 Guidelines vol. 2 for fuels. Printing defaults: Kujanpää and Nors 2014, Environmental performance of future digital textile printing, VTT-CR-04462-14. Their values, versions, and sources are written on the dashboard and in reports beside the figures that use them.

Rule version history

What changed, when, and what did not change with it. A released report is frozen at the versions of its release.

13 August 2026: factor library v1.0-provisional, the first.
4 September 2026: factor library v2.0-idemat-2026RevB1 replaces v1.0-provisional.
24 September 2026: factor library v2.1-idemat-2026RevB1 replaces v2.0-idemat-2026RevB1.
23 September 2026: Circular potential rules v2: fibres count by family, an inherently recycled fibre counts as 100% on its line, a documented return route earns points, and a product without a composition gets no score. Rules v1 stopped applying the same day.
23 September 2026: brand-level figures changed from a sum of one unit of each product to a simple average per unit with a coverage line. Released reports did not move; their label names the old form.
24 September 2026: the lifespan estimate calculated by Mula, in force for one day from 23 September 2026, was withdrawn entirely and appears on no surface. What shows since then is the brand's statement, exactly as written.

What is not scored

No letter grade on the passport. We calculate a materials grade for the brand inside its own dashboard, so the brand knows where it stands and what would move it. That letter never appears on a public passport. The reason: the rest of the passport states facts that can be checked, while a letter ranks a brand against a scale we drew up ourselves, and that scale would stand on the non-Indonesian factors we just described. If you see a letter grade on another platform's passport, it is worth asking which factors it used and where its band boundaries are published.
Certification does not set the value. A certificate is evidence, not virtue. Scoring it would penalise exactly the upstream that cannot afford the gate.
Origin and import distance do not set the value. Transport is a fact that belongs in the carbon footprint, not a judgement about a place.
These figures are not third-party verified. Our calculation calculates; it does not certify. A figure becomes verified only through third-party review. Where a brand uploads an LCA study that has been through third-party review, the study's figure replaces the estimate on the metric it covers, with the assessor's name and year on its card.

Address and resolution

The figures can change. The address cannot.

The figures on a passport can change as its evidence improves. The passport's address cannot. This section writes down the address rules, because the address printed on a hangtag is the one part of this system we cannot correct after it ships.

The address form

Every passport lives at p.mula.asia/{brand-code}/{slug}. That subdomain is used for nothing else: not marketing, not the application, not redirects. Marketing routes may be rearranged; passport routes may not.

A slug is editable until first publication and permanent thereafter. An address that can still be changed after it is printed is a broken QR waiting to happen. Slugs are unique within each brand's namespace and checked against a reserved-word list.

The address also does not change when the state of the passport changes. A passport that goes inactive, that becomes active again, or that moves to a different plan stays at the same address. What changes is what renders there, never where the code points.

The address is deliberately short. Every character adds modules to its QR, which densifies the code, lowers legibility on a small tag, and leaves less room for the mark at the centre.

One printed form, always

Only the path above is ever printed. The canonical path /01/{GTIN} is answered at the same address as a redirect alias, from day one, for machines: customs systems, retailer systems, recycler scanners, and registry lookups. Those systems query the canonical path and do not scan hangtags.

That redirect is temporary rather than permanent, and that is deliberate. The passport address is permanent, while the GS1 number pointing at it is data a brand can edit. A permanent redirect would sit in browser and proxy caches longer than the fact behind it survives, and a cache cannot be recalled any more than a printed tag can. The canonical path is also kept out of search indexes, so that it does not slowly become a second address.

A number with a wrong check digit and a number that points at no passport are answered differently. The first is a malformed number, the second is a question with no answer, and the system on the other side needs to know which one it has to fix.

A brand that gives each individual piece its own identity uses the same address with one more level of depth, p.mula.asia/{brand-code}/{slug}/21/{serial}. That is one address form with two depths, not a second scheme.

Three reasons the GTIN path is never printed

A GTIN is per variant, while a passport is per product. A shirt in five sizes carries five GTINs against one passport. Printing the GTIN path forces a choice between five passports and one number that is factually wrong on four tags out of five.
Ownership of a number cannot be verified. The check digit proves the number is well formed, not that it belongs to the brand using it. A typo becomes permanent at print.
Two printed schemes in the world means two code paths and two failure modes, against a constraint that is really working hours.

Product number (GTIN)

Optional, at variant level. The check digit is validated on entry. Never required, never a gate on publication.

Mula never issues a number and offers no control that generates one. An invented thirteen-digit number will eventually collide with a real product, and by then it is printed.

When present, the GTIN renders small in the product details line beside the SKU, and always travels in machine-readable output and in the export pack. When absent, nothing renders: no empty field, no note saying the number is missing, no gap indicator. A brand without GS1 membership is not an incomplete brand, and rendering its absence would let a paid credential back into the passport through the side door.

Where we stand with GS1

Mula passport addresses follow the GS1 Digital Link URI structure.

That is the only form of the statement we use, and we do not use a stronger one. Mula is not a GS1 member. Mula is not GS1 certified, not declared conformant by GS1, not registered with GS1, and not a GS1 partner. We do not display the GS1 logo. Following the structure of an address is a technical choice; membership and conformance programmes are something else, and we are not in them.

Brands using Mula do not need to be GS1 members either.

The European Union registry

The EU DPP registry has been running since July 2026. It is a directory that indexes identifiers and points at passport data; it is not a data host. No product registration duty falls on textile brands until the textiles delegated act triggers it. When it does, registration requires a verified economic operator, which for Indonesian brands runs through their EU importer or authorised representative, not through Mula.

Registered Indonesian private-scope electronic system operator (PSE) No. 029287.01/DJAI.PSE/08/2026