{"positioning":"Commodex earns institutional trust through transparency: platform-level activity, token backing and governance — every figure a real query at request time, never a marketing number, never identifying any participant.","as_of":1790511494,"platform":{"data_provenance":{"figures_derived_from":"the deal book, bucketed by commodity class","book_origin":"UNVERIFIED ORIGIN","seeded_deals":null,"post_seed_deals":null,"rails":"no payment rail on this platform is live — settlement rails, custody and bank facilities are external and recorded as pending — so a deal originated through the product is an originated record, never evidence that money moved","doctrine":"a figure published without its origin is the honest half of a misrepresentation — this marker travels with the number"},"value_transacted":{"buckets":[{"commodity_class":"gold","deals":10,"value_usd":1210000000.0},{"commodity_class":"other (k-anonymity merge)","deals":10,"value_usd":1190000000.0}],"deals_suppressed":"none","deals_suppressed_max":4,"value_suppressed_max_usd":0.0,"classes_suppressed":0,"k_anonymity":"a class with fewer than 5 deals is merged into 'other', and 'other' is itself published only once it reaches 5 deals — below that it is withheld entirely, so no SINGLE READING of a public bucket can be resolved to one transaction. `deals_suppressed` is a BAND, not a count — an exact integer of withheld deals moved by one every time a deal in a thin class settled, which is a per-record signal about the very records this control protects; `deals_suppressed_max` is the machine-readable bound, and the published bucket counts are quantised for the same reason. WHAT A WITHHELD CLASS COSTS THE READER, IN THE UNIT THIS BLOCK IS KEYED IN: a suppressed class takes its VALUE out of the published buckets too, so `value_suppressed_max_usd` bounds that withheld value and `classes_suppressed` says how many classes the bound covers. It is a DECADE ceiling, not a total — coarser than the published figures on purpose, because the withheld population can be a single deal and a tight bound on one deal is that deal. HOW TO ADD THESE UP, SINCE THE OBVIOUS ARITHMETIC IS WRONG: every published bucket count is rounded DOWN to a multiple of 5, so sum(buckets.deals) is a LOWER bound on the settled book and sum(buckets.deals) + deals_suppressed_max is NOT an upper one — each bucket can already be understating by up to 4. Bucket VALUES are quantised to three significant figures in either direction, so their sum is an approximation of the published classes rather than a bound on them; the only strict money bound published here is value_suppressed_max_usd, over the withheld classes alone. WHAT THIS DOES NOT COVER, STATED RATHER THAN IMPLIED: every bucket here is a LIVE running total recomputed on each request, so the DIFFERENCE between two readings taken either side of one settlement BOUNDS that settlement rather than naming it — the published value and deal count are both quantised, so the delta is a bound and no longer the exact contract value: measured anonymously on this endpoint, a $5,760,000.00 settlement moved a published bucket $124,000,000 to $130,000,000 with the published count unchanged, and a settlement that crosses a quantisation step still moves both. k-anonymity bounds a snapshot, not a time series; treat the published totals as disclosed-by-bound and do not read this sentence as a promise that a polling observer learns nothing about an individual trade"},"participants_by_category":[{"category":"compliance","count":6},{"category":"admin","count":1},{"category":"insurer","count":1},{"category":"producer","count":1},{"category":"merxgold","count":1},{"category":"logistics","count":1},{"category":"sovereign","count":1},{"category":"auditor","count":1},{"category":"institution","count":1},{"category":"financier","count":1}],"corridors_registered":90,"certification_councils":8,"certifications_issued":2,"disputes_resolved":0,"disputes_escalated":0,"disputes_open":0,"listings_open":0},"token_backing":{"note":"CGX discipline: 1:1 allocated grams — issued may NEVER exceed vaulted. Verify the raw records yourself: /api/reserve/attestations (signed attestation history) and /api/chain/status (the tamper-evident evidence chain, with the coverage it claims).","attestation":{"seq":2,"at":1790510531.1267262,"vaulted_grams":18498.15,"issued_grams":9000.0,"backing_pct":149.98},"fully_backed":true,"evidence_chain_intact":true,"evidence_chain_coverage":{"verified_through_event":152,"events_in_store":232,"unsealed_tail":80,"unanchored_inside_sealed_range":0,"note":"verified_through_event is the last event a sealed Merkle root actually covers. Anything above it exists but is not yet anchored, and this verdict makes no claim about it."}},"clarity":{"thesis":"The single most historically leaky step in commodity trade — moving resource revenue from the point of export to the treasury — is what this register exists to make checkable: every corridor's royalty rate, authority and legal basis is published here, with the accrual computed on the deal value the cited code actually levies, so what is owed can be compared against what is paid.","royalty_priority_status":{"claimed":"royalty tranche released to the named sovereign account before anything else","ordering_built":true,"movement_built":false,"applies_to":"deals that use the OPTIONAL paymaster structure — one settlement option among several. Most deals settle another way (documentary settlement, escrow release on outturn, documentary credit, offset, netting) and never enter a waterfall at all. This flag describes what the structure enforces when it is used, not what every deal on the platform does.","reality":"ORDERED AND HELD, NOT YET MOVED. When a deal uses it, paymaster.py computes and records the disbursement instruction with the sovereign royalty at rank 1 and Commodex's own margin at rank 6; the ordering is a code constant, every beneficiary is screened before it is instructed, and a leg that escrow cannot cover is recorded UNFUNDED rather than part-paid. AND THE ESCROW NOW HOLDS THAT MONEY: an instructed leg encumbers the balance it is drawn against, so the escrow cannot be released elsewhere while the royalty stands unpaid — until 2026-08-15 the ordering was real but the money was not held, and the whole balance could be released out from under an instructed royalty which was then confirmed anyway. The transfer itself is still NOT executed here: no settlement rail is authorized, so each leg stays 'Instructed' until an operator confirms it against a real rail reference, and confirming it debits this platform's own escrow record rather than moving funds. A live rail and a sovereign mandate remain EXTERNAL_BY_DESIGN.","ordering":["Sovereign royalty / export levy","Logistics and insurance","Trade-finance repayment","Producer / supplier","Documented intermediary commission","Commodex net margin","Certification fee"]},"countries":[],"suppressed_below_k":2,"honest_caveat":"What this register makes verifiable is the ACCRUAL: each corridor's rate, authority, legal basis and the deal value that basis actually levies, published so what is owed can be checked against what is paid, and the ORDER in which it is instructed: the royalty ranks first and Commodex's own margin ranks sixth, fixed in code. It does NOT move the money — no settlement rail is authorized, so a leg is an instruction until someone confirms a real transfer against it (see royalty_priority_status). What a government does with revenue inside its own treasury is a domestic governance question no commercial platform can honestly claim to solve. The precise claim is already unprecedented for this industry; it is never overstated into 'we end corruption'."},"intelligence":{"sources":{"intel_engine":{"total":32,"live":2,"licensed_never_returned_data":1,"licensed_endpoint_unreachable":7,"licensed_reachable_no_matches_yet":0,"licensed":10,"pending_credential":6,"pending_contract":16},"atlas_universe":{"providers":90,"live":6,"pending":84,"by_category":{"exchanges":11,"vessel_ais":6,"aircraft":3,"weather":7,"satellite":8,"altdata":1,"energy":7,"finance":5,"battery":2,"trade_flow":5,"ports":4,"freight":3,"insurance":3,"alt_data":5,"platform":1,"prices":1,"agri_grain":2,"agri_soft":1,"precious":4,"geopolitical":2,"seismic":1,"rates":1,"trade":1,"base":1,"bulk":1,"news":2,"shipping":1,"sanctions":1},"by_status":{"pending_contract":43,"pending_integration":20,"pending_credential":13,"live":6,"open_licence":8}},"reliability_scale":"1-5★ spine with letter equivalence (5★=A+ … 1★=D); LIVE only on a real feed record"},"forecast_ledger":{"recorded":{"matured_this_run":0,"pending":12,"scored":{},"hit_rate":null,"realized_series_frozen":[],"honest_note":"hit_rate is measured against realized prints only; None until forecasts mature."},"matured_scoring":{"pending":12,"scored":{},"hit_rate":null},"honesty":"hit_rate is measured against realized prints only; None until forecasts mature."},"corrections":"the forecast-outcome ledger above is the live public accuracy record — every forecast is scored on maturity, hit and miss alike. A separate source-revision log is NOT BUILT: there is no corrections store in this platform, and this register will not claim one until there is."},"governance":{"councils":["CCA","CCB","CCC","CCD","CCE","CCEC","CCGC","CCMC"],"structure":"a foundation-anchored group — operating entities are disclosed at entity-role level on authorized surfaces; sister ventures earn visibility, never privilege (marketplace-neutrality doctrine). The control architecture is not a public fact."},"anonymization":"Platform aggregates only. No organisation, party pair or deal reference is ever published here; buckets respect k-anonymity; the Momentum feed applies delay, banding and regional generalisation on top — every entry is deferred 6-18 minutes before it can appear, sizes are published as bands and never as figures, and a corridor is generalised to region level when naming it would break k-anonymity. The full anonymization register is an authorized-tier surface (/api/atlas/anonymization), not a public one."}