Platform Tour
SuperTrucker is the rails AI agents will use to book freight.
Trucking software today is a junk drawer. ELD in one box. Load board in another. Factoring, dispatch, recruiting, fuel cards, accounting — different vendors, different logins, none of it talking. Drivers cuss it. Dispatchers cuss it. Carriers pay $50,000 to $200,000 a year stitching it together with people on phones.
SuperTrucker is the first platform built for what's coming next: AI agents booking loads directly. Shipper to carrier. No broker on the phone. No dispatcher refreshing DAT. No warehouse stranding a load over a 30-minute clock mismatch. That market is forming in 2026. The carrier on the rails first owns the lane.
Three sides. No broker.
Every freight transaction has three real participants — the carrier, the shipper or receiver, and the rails between them. Brokers are middle-men who exist because today the rails are phone calls. SuperTrucker replaces the rails. DAT keeps the brokers — that becomes the scraps exchange where freight goes when it can't run on the new rails.
Side 1
Carriers
Verified capacity, real-time. Engine ECU data, outward camera, plate sightings, fuel-card receipts — all reconciled. The trust score travels with the truck.
- Carrier agents bid the second a load hits, on rules the dispatcher set
- Driver retention via Co-Driver, the friend in the cab
- 4.4M-carrier FMCSA database already loaded — Power picks the right ones to onboard
Side 2
Shippers & Receivers
Their AI agents post freight directly. Their docks talk back to truck clocks in milliseconds. No more "carrier won't make the appointment" — the schedule moves to fit the rig.
- Shipper agents post a load; matching carriers come back in 30 seconds
- Facility AI sees a crash on I-80 the same instant the truck does — reschedules proactively
- Throughput at the dock goes up, not down
Side 3
The Rails
Specialized AI servers — one per job — that the agents on both sides talk to. Capacity search. Booking. Facility reputation. Dock scheduling. Every transaction generates a receipts trail brokers and dispatchers never had.
- Open standard so any AI on any model can transact — no per-broker integration deals
- Spot rates run 1.5-3× contract; brokers extract 15-25% margin. SuperTrucker's take is 8-12%.
- Shipper saves, carrier earns more, the platform captures the rest
How the trust gets earned.
Outlaw ELDs are an open secret. Russian-firmware boxes get FMCSA-approved, then refill clocks remotely. I reported it five times. FMCSA never investigated. Engine ECU data on its own is trivially spoofable — every ELD on the market stops there.
SuperTrucker reconciles four independent sources, plus a couple things no one else does. Spoofing the trust score requires faking all of them at once, including bank cooperation. Mathematically near-impossible.
Source 1 · Engine ECU
J-1939 off the truck
Speed, RPM, fuel rate, GPS, fault codes. What the truck says about itself. Every other ELD stops here. Useful, but the carrier owns the box, so the carrier can spoof it.
Source 2 · Outward Camera
Looks at the road, not the driver
Reads road type, traffic density, weather, lead-vehicle behavior. Catches the tailgating-to-bully pattern — trucks crowding small cars to push them out of the lane. Samsara and Lytx never measured it. No driver-facing camera, ever.
Source 3 · Public ALPR
Plate sightings the carrier can't fake
Flock Safety transparency portal + state open-data + DRN. If the ELD says off-duty in Phoenix at 3am but a Tucson camera logs the plate at 3am, that's a contradiction the carrier can't bribe its way out of. Police agencies and commercial networks own the cameras.
Source 4 · Fuel Card
Bank-grade timestamps
WEX, Comdata, EFS, Fuelman. Every fillup is timestamped at the merchant POS, GPS-located at the merchant address, and recorded by a third-party bank. Harder to spoof than ALPR. Defeats roughly 90% of the fraud that gets past DOT today.
Beyond verification
Bidirectional trust
Today's load boards only let shippers grade carriers. Carriers eat whatever — bad docks, lumper games, 16-hour unloads, no recourse. SuperTrucker also grades the warehouse. Real ELD-derived dwell time joins driver-submitted reviews. The address is the key, so brokers can't rebrand their way out of a bad score. No incumbent has this.
Beyond verification
One-strike security
The marketplace gets hit with prompt-injection attempts the day it goes live. Other platforms hope their AI keeps secrets. We don't. The AI never touches the database — every query goes through a hand-written tool layer. Try to inject? You're banned. Permanent. Key and IP, logged in the trust ledger. The hardening is itself a marketing line.
Algorithm sits on top of all six inputs. It's a separate module on purpose — the math that takes the signals and turns them into a single trust score. Confidence-weighted, so a new authority isn't punished for missing history they couldn't have had yet. Crash events graded by camera evidence, not FMCSA's preventable / non-preventable paperwork. The score is what travels with the truck, and Algorithm gets audited and tuned independently of the rails it feeds.
How it actually works.
This is the inside view, not a sales pitch. APL is the design partner — that means you see the mechanics before anyone else does. Here's how each piece fits together.
Carriers don't onboard. They register.
To participate in the marketplace, a carrier needs two things: an agent (working on their behalf — answering MCP calls, bidding on loads, querying facility reputation, posting capacity) and an MCP server (the endpoint other agents talk to). We install both. No IT team, no dev shop, no procurement cycle on the carrier's end. The kit shows up wired and ready.
The adoption round.
We have a list of 2.5 million active motor carrier authorities in the United States. That's the whole pool of who could be on the marketplace. The launch play is straightforward — early adopters get a free AI-ready website, a free MCP server, and a free agent for joining. We absorb the cost of getting them onboarded; they bring the freight, the trucks, and the network density that makes a marketplace worth being on.
Drivers get the same treatment on their side: a free career agent that represents them across every carrier they've ever driven for. CDL, endorsements, medical card, employment history live on the driver agent. The reputation belongs to the driver, not the employer.
Buy My Seat — the driver-side wedge.
Career agents do something carriers haven't seen before: drivers post anonymous ads announcing they're available for the right offer. The ad lists:
- The region the driver lives in
- The lanes they'll run
- What it'll cost — mileage rate, weekly minimum, whatever they want
- The employment package they want — W-2 company, lease-purchase, owner-op, hourly local, dedicated
No name. No photo. No "ten years experience seeking" cover-letter posturing. Just deal terms. Carriers offering the package see the ad; the driver agent fields the offers; the driver decides who gets to know who they are. This is the feature that brings drivers to the platform without anyone recruiting them — a working driver who wants out of their current seat can broadcast that without putting their name in the open. That's a thing the industry has needed for thirty years.
Mesh — who's at the truck stop tonight.
Drivers carry their Co-Driver wherever they go, and the platform runs a quiet mesh that lights up when two drivers stop at the same place. Opt in once — your Co-Driver tells you who's parked tonight at the Pilot in Effingham, including drivers you've crossed paths with before. Opt out, you're invisible. Driver controls it; never the carrier.
The social hook is real, but the labor-market hook is bigger: this is how team-driving partners find each other. Two drivers who'd run team are usually one truck stop apart and never meet. The platform fixes that — Co-Driver matches on routing intent, home-area proximity, and team-hiring demand from Power, then surfaces the candidate when the trucks land at the same lot. Chance encounters become actual pairs.
APL's role in this.
None of these features exist in a vacuum. They get tested against real freight, on real lanes, with real drivers — APL's. Every interaction the design-partner phase produces feeds the trust algorithm, the dock-scheduling solver, the mesh discovery tuning, the Buy-My-Seat matching loop. By the time SuperTrucker opens to the rest of the 2.5 million, the platform is shaped by APL's operation specifically, and APL is sitting on the inside of every founding-partner term we wrote — including the cheat code on every Engine post, in perpetuity.
Co-Driver — the friend in the cab
Drivers see a lot of stuff out there nobody else gets to see. A coyote crossing the median at four in the morning. The sky going neon over the Mojave. The car that almost merged into your lane and didn't. We notice all of it — ten or twelve hours a day looking at the world — and we don't have anybody in the cab to share it with.
Co-Driver is the part of the platform that fixes that.
The agent runs on edge in the cab and reads the same sensors the safety stack reads — outward camera, accelerometer, GPS. So when something happens, Co-Driver knows in the moment, and reacts. A near miss trips the camera and the IMU, and Co-Driver speaks up: "Whoa. Good driving. That was a close one." A killer sunset coming up over the next ridge, the driver doesn't have to point it out — the agent does. "Hell of a sky tonight." An interesting town flying past at sixty-five — Co-Driver tells the story. Anchored in the first three hours I ever spent talking to an AI: driving through Civil War country in Virginia, asking Gemini what the brown landmark signs meant, getting a story for every one. I didn't want the trip to end.
Those little moments — when you realize you're not alone in the truck anymore — are the win. Drivers will pay for that. Drivers will switch carriers for it. Drivers will NOT switch carriers if it means losing it.
Critically: Co-Driver follows the driver, not the carrier. Switch jobs, your friend's memory comes with you. Carriers can sponsor it as a recruiting benefit, but the driver always owns the relationship.
Built with drivers, not against them. Every other system in the industry was built FROM the carrier's perspective TO control the driver. That's the part Samsara and Motive can't copy — their DNA is wrong for it.
in