A shipment isn't a row in a broker's database. It's one permanent address that the shipper, the carrier, the broker, and the receiver each open with their own key, seeing exactly the slice they're entitled to. Costs are visible, escrow is funded before the truck rolls, and the record can't be quietly edited.
Every party on a load gets a view built for their job, and a page here that explains what changes for them.
Every load today is two sets of paperwork, a broker's private ledger, and 30 to 90 days of waiting. The tools built on top of that are serious. The foundation is the problem.
The broker owns the data. Margins are hidden; the carrier and the shipper each take the broker's word for the rate.
Net-30 to net-90, or you factor and lose 3–5% to bridge the gap. The money exists; it just isn't yours yet.
Identity isn't bound to the load. A different truck shows up, or the load is quietly resold to a carrier nobody vetted.
Rate con, BOL, POD, invoice: four documents, four versions of events. Whose paperwork is right?
A load is created, funded, accepted, picked up, moved, delivered, and settled. Only the right party can move it to the next state, and each step is appended to the record, never overwritten.
Your real USDOT, MC, or EIN becomes your address. No new account number, nothing to re-issue when FMCSA finishes moving to USDOT-only identifiers.
Terms, parties, tracking, and settlement all live at it. The shipper funds escrow, the carrier binds a truck and driver, and each party sees only their slice.
Verified pickup releases 30% to the carrier. The receiver's confirmation releases the remaining 70%. The record seals.
Walk through the full lifecycle, with who does what and what money moves at each step →
Load Guardrails is the record those things run on. Four things follow from building the record this way.
Driver qualification files, equipment, insurance, and settlements live in the carrier's own field under the carrier's key, not in a broker's or a vendor's database. A carrier that changes brokers keeps its whole operation.
Rate con, bill of lading, proof of delivery, and invoice are entries on one address every party already reads. There is no 204-to-990-to-214-to-210 loop to reconcile, because it is the same object.
A load is bound to a real USDOT and a specific truck and driver at acceptance. Re-brokering it is not a risk to monitor; it is a state change the policy refuses to write.
A load can't be tendered to a carrier whose authority, insurance, and packet aren't active and adequate. The check runs before tender, every time, and it is deny-by-default.
Unbroker runs its carrier network, vetting, and loads on Load Guardrails. It publishes its fee, drops the non-solicitation clause, and when a shipper and carrier want to go direct, it steps aside with seven days' notice. That switch is a recorded, consented operation on the load, not a lawsuit.
unbroker.io → How brokers and platforms build on Load Guardrails →
Everything here is ordinary technology used well: a shared record, keyed access, an append-only log, and prepaid escrow. We don't claim a system can do what systems can't. Freight doesn't need magic; it needs one honest record every party can trust, and that's what this is.
Carriers, shippers, brokers, and platforms: get on the interest list and we'll start the onboarding conversation ahead of launch.
Join the interest list