A Startup Uses Existing Security Cameras to Measure Freight—No Special Hardware Needed

A San Francisco startup called Transload is building software that extracts freight dimension data from security cameras logistics operators have already deployed — no dedicated dimensioning hardware required. The company emerged from Y Combinator's P26 batch in 2026 with a three-person team and a narrowly focused problem that turns out to be surprisingly widespread across warehouses and freight docks worldwide.

The Problem Transload Is Solving

Accurate dimensional data — length, width, height, and weight based on volume — is essential to freight pricing, load planning, and carrier compliance. Yet the industry has historically generated that data in one of two ways: manually, with workers using tape measures and laser tools, or through purpose-built dimensioning systems that require dedicated camera rigs and integration work that can cost well into six figures before the first package is measured.

Neither approach works well at scale. Manual measurement is slow, inconsistent, and creates bottlenecks during busy periods. Dedicated systems are expensive for mid-market shipping companies and regional carriers, and they typically only cover the specific spots where they are installed — leaving the rest of the facility essentially unmeasured. The result is a chronic gap between stated dimensions and actual freight size, which creates errors in rate calculations, load plans, and billing disputes.

Transload's idea is to close that gap using the infrastructure that is already there.

How It Works

The company applies computer vision — software that can interpret images — to the video feeds from standard security cameras already mounted in docks, staging areas, and warehouse floors. According to its Y Combinator profile, the system analyzes freight moving through these environments during normal operations — no special scanning workflow, no additional hardware installation, no need to move or reconfigure existing cameras beyond what the software can handle.

The technical challenge is real. Security cameras are built for surveillance — they have wide-angle lenses, compressed video, variable lighting, and mounting positions chosen for coverage rather than precision measurement. Getting reliable three-dimensional measurements from that kind of footage requires solving for lens distortion, partial occlusion (when freight is partially hidden), perspective shifts, and the absence of depth sensors that dedicated dimensioners rely on.

The software almost certainly combines two approaches: estimating depth from a single camera view (monocular depth estimation) or calculating 3D position from overlapping camera fields, anchored to known reference objects for scale. Getting measurement error tight enough for shipping billing — where a few centimetres can shift a shipment into a different pricing tier — is exactly the kind of problem that has become solvable as large vision models have become more capable and the computational cost has dropped into commercially viable range.

The Team and the Founding Context

Transload was founded by Nils Börner, Julius Scheel, and Jago Wahl-Schwentker. At three employees, the company operates at classic early-stage density: the founding team is essentially the entire engineering, sales, and product organization. Y Combinator's P26 batch places the company's formation in early-to-mid 2026, meaning the product is likely in active pilot or early customer deployment rather than operating at full scale.

The Y Combinator backing matters beyond just capital. The accelerator's network provides access to logistics operators and enterprise contacts that would otherwise take years to develop cold, which is valuable for a small team selling into an industry where vendor trust and proof of financial return move slowly.

A Familiar Pattern in Industrial Computer Vision

We have seen this playbook before. When machine vision first appeared in factories in the 1990s, the dominant model was purpose-built: dedicated camera systems, custom lighting rigs, bespoke integration work. Over time, cheaper image sensors and better general-purpose vision software gradually made dedicated systems less necessary. Once smartphones put high-resolution cameras in everyone's pocket, the logic flipped entirely — the question stopped being "how do we install cameras?" and became "how do we extract more value from cameras already in place?"

Transload is applying that same logic to logistics facilities. The installed base of IP cameras (networked security cameras) in commercial buildings is enormous and mostly used only for passive recording and security monitoring. Treating that installed base as a sensor network for operational data — rather than just a security asset — is a thesis that other startups have pursued in adjacent areas, from retail inventory tracking to manufacturing quality control. Freight dimensioning is a logical next target: the financial incentive is clear (better billing accuracy and load efficiency), the existing measurement gap is well-documented, and the customer adoption barrier is low because no new hardware is required.

What the Market Context Looks Like

The logistics technology industry has been modernizing unevenly. Warehouse management systems and transportation management systems have received significant investment over the past decade; last-mile routing has been heavily optimized; automated sortage is standard at large facilities. But the physical measurement layer — actually knowing what you have, in three dimensions, as it moves through your facility — remains incomplete outside the largest companies.

Dimensional weight billing, now standard across major parcel and freight carriers, has created real financial incentive for accuracy. Shippers who declare dimensions that are too small face correction surcharges; carriers who fail to capture correct dimensions leave money on the table. The industry estimates the annual gap in billions of dollars, though exact figures vary by segment and geography.

A software-only approach that works with existing camera infrastructure directly addresses the adoption barrier that has kept dedicated dimensioning out of mid-market facilities. If Transload's accuracy holds up across the real-world variability of active docks — inconsistent lighting, forklift blockages, mixed freight types — the sales friction is substantially lower than any hardware-dependent competitor.

What to Watch

The critical unknowns at this stage are measurement accuracy across diverse real-world conditions, the integration pathway into existing warehouse and shipping software workflows, and whether the team can close enough early customers to establish the reference case that enterprise logistics buyers require before committing.

The broader context here is worth noting: computer vision accuracy in controlled pilot environments does not always hold up when it meets the full complexity of an active freight dock — variable lighting, weather-opened dock doors, partial loads, and the sheer diversity of packaging formats. How Transload handles edge cases, and what accuracy guarantees the company can offer in commercial contracts, will be the real test of whether the technology is ready for production use.

The broader direction, though, is clear. Extracting operational intelligence from infrastructure already in place — rather than layering new sensors on top — is a durable trend in industrial software. Transload is a small, early bet on one specific node of that transition. Whether this particular team and product become the reference solution for freight dimensioning, or serve as a proof of concept that a better-capitalised competitor eventually absorbs, the underlying thesis is sound.

---

Sources: Y Combinator — Transload company profile; Y Combinator — San Francisco Bay Area companies.