Transport capacity is the purest perishable good in commerce: a vessel open in a port on Tuesday, or a truck running back empty, is worth something on Monday and nothing on Wednesday. It is also priced almost entirely by conversation, against indices built from what brokers believe the market to be rather than from what it actually did.
Chartering works, at scale, and has done for a very long time. It also concentrates almost all of the market's information in the brokers who sit between the two sides, which is fine until you are the party that does not have three of them on speed dial.
Freight benchmarks are built from panel submissions: brokers reporting where they believe the market sits for a notional voyage. They are carefully constructed and genuinely useful, and they are not transaction prices. Nobody can point to the trade that made the number.
Vessel class, laycan window, route, cargo, demurrage rate, bunker clause and off-hire terms are negotiated per fixture. Two fixtures at the same headline rate can be materially different deals, which makes the rate almost useless as a reference for the third.
A broker ringing round to place an open vessel is racing a clock that ends in a ballast leg. The urgency is real and it is asymmetric — the party that knows how little time is left extracts the value of that knowledge.
Commission on conclusion is a reasonable model and it does not reward publishing the price. There is no villain here and there is also no mechanism by which the market becomes more transparent on its own.
The trick is separating the unit from the contract. The charter party can stay as bespoke as it has always been; what has to be standard is the thing whose price is being discovered.
A defined voyage — this route, this vessel class, loading in this week — is a book. The commercial terms that genuinely vary stay in the charter party the parties sign afterwards. The market prices a standard unit; the lawyers keep the bespoke document.
A daily uncrossing at a published time gathers a whole day of scattered interest into one moment and clears it at one price. For capacity that is worthless after a date, a scheduled call is a better mechanism than a race between telephone calls, and it does not reward being the last person told.
The uncrossing price of a defined instrument is a transaction, formed from live orders. Panel assessments become a sanity check on the market rather than the only description of it, and the rest of the chain — including the physical contracts that never touch the venue — has something to quote against.
GTD validity lets an owner leave an offer working through several days' auctions as the open date approaches, and lets a charterer keep a bid live while cargo readiness firms up. The market runs at the speed the underlying business actually moves.
An iceberg order displays a peak and hides the remainder while contributing its whole volume to the uncrossing. An owner with several ships open in the same region can bring them to market without announcing exactly how exposed they are.
Scheduled auctions on a published clock, credit as a pre-trade limit rather than a phone call, and netting across a day of fixtures.
| Setting | Typical starting point |
|---|---|
| Instrument | Route × vessel or vehicle class × laycan week |
| Lot size | One voyage, a fixed tonnage parcel, or one vehicle-day |
| Price | Currency per tonne, per day, or as a differential to a published index |
| Sessions | One call auction a day at a published time per route |
| Validity | DAY and GTD, so an order rests across the laycan window |
| Risk | A named profile per member, replacing the bilateral credit call |
| Settlement | Netted per participant per date; the voyage performed outside the venue |
One call auction per instrument per day with per-day overrides for holidays and seasonal patterns, plus continuous trading on the routes liquid enough to carry it.
Read more →04 · coreLimits at clearing-member, trading-member and end-client level: maximum order size, notional and open position. This is what replaces the question of whether these two firms have a line with each other.
Read more →05 · post-tradeMultilateral netting of freight obligations per participant and settlement date, fees per fixture, and enriched trade-capture records for the operations and post-fixture teams.
Read more →06 · channelHTTP/JSON with server-streamed events for the chartering desks and brokers' own systems, and snapshot-then-stream market data so a rate screen is the book rather than a poll of a cache.
Read more →Mehrex matches a rate for a standard unit. Everything that makes freight physical — the contract, the vessel, the port, the cargo and what happens when any of them fails — stays outside it.
A matched order is an agreement to fix at a price. The instrument that binds the parties is a charter party or haulage contract with laytime, demurrage, off-hire, bunker and cancellation terms, under a governing law with an arbitration seat. Standard forms exist; getting your market onto one is a commercial project.
Charterers vet vessels; shippers vet hauliers. Age, class, flag, inspection history, insurance and cargo compatibility all gate whether a nominated carrier is acceptable at all. That gate has to sit in membership and instrument eligibility, and someone has to operate it.
The common failure is not non-payment but non-performance: a vessel that does not arrive in the laycan window, a load that is not ready. Pre-trade limits do nothing about that. Substitution rules, cancellation rights and a remedy framework are rulebook work.
Port agency, bunkering, scheduling, documentation, customs and demurrage claims. The trade-capture record tells your operations team what was fixed; running it is the business you are actually in, and it is where the staff cost sits.
Broker relationships are the liquidity. A venue that positions itself as their replacement launches into an empty book with a hostile industry; one that gives them a member account, an execution channel and a defensible price has a chance. This is the commercial decision the whole project turns on.
Physical fixtures are one thing; a standardised, transferable, index-referenced freight contract is treated as a derivative in many jurisdictions, with venue authorisation and reporting attached. Freight derivatives already trade on that basis. Establish which side of the line your product sits on before you design it.
It does not — the two are separated. The instrument standardises only what the price refers to: route, vessel or vehicle class, and laycan window. The parties still sign a charter party covering the terms that genuinely vary between them. The market discovers a rate; the contract governs the voyage.
Because interest in a specific route and laycan is scattered across a day and the good expires. A daily uncrossing concentrates that interest into one moment and clears everyone at one price, which produces a more reliable reference than a continuous book that is empty most of the time. Mehrex runs either, configured per instrument.
It should not, and a venue designed on that premise will fail. Brokers hold the relationships and the cargo knowledge that make a market liquid. The realistic model is brokers as members, executing on a venue that gives their clients a defensible price rather than an assessment.
Accelerator hours and colocation power — extremely perishable, priced bilaterally, and hoarded because there is no spot price.
Read the page →Capacity & accessSecondary trading and leasing of licensed spectrum and wholesale interconnect capacity — permitted in principle, almost nonexistent in practice.
Read the page →Precious metalsAn app that sells sub-gram quantities of gold to retail buyers, run as a real order book instead of a dealer’s quote.
Read the page →A live demonstration runs about an hour: the order book, the Control Panel, an auction uncrossing, and a replay of the journal that produced it.