Damm Intelligent Logistics System

Inspiration

Damm Smart Truck was created as a logistics prototype for beverage distribution, focused on improving how warehouse teams prepare pallets and how drivers execute deliveries.

The idea came from a real operational problem: in multi-stop delivery routes, a poorly prepared vehicle can force the driver to search through too many pallets, move heavy items unnecessarily, or deal with returns without enough reserved space. This is especially relevant when the load includes mixed products such as beer barrels, returnable boxes, water, glass bottles, food products and cleaning supplies.

What it does

Damm Smart Truck simulates an intelligent loading and delivery assistant for warehouse and driver workflows.

The application includes two main roles:

Warehouse: prepares the vehicle, reviews the loading plan and checks how pallets should be organized.
Driver: selects the assigned vehicle, sees pending stops, opens delivery notes and follows unloading instructions.

The prototype includes:

a vehicle selection screen with realistic load occupation,
optimized pallet assignment,
stop-by-stop delivery workflow,
a real PDF delivery note integrated into the application,
structured interpretation of the delivery note,
logistics calculations for boxes, barrels, SKUs, weight and estimated volume,
pallet position information,
warnings for heavy or delicate products,
a voice notification system with browser demo mode and optional ElevenLabs integration for real usage.

How we built it

The project was built as a Vite + React + TypeScript application.

The demo uses structured mock data to simulate vehicles, stops, delivery notes, pallet assignments and route logic. A real delivery note PDF was added to the project and linked to a route stop, allowing the application to show both the original document and an interpreted logistics summary.

The voice system was designed with two modes:

browser mode for the demo, using native speech synthesis without consuming external credits,
real mode using a secure backend proxy for ElevenLabs, keeping the API key outside the frontend.

The backend proxy was implemented with Express, so the frontend never calls ElevenLabs directly and does not expose private API keys.

Challenges we ran into

One of the main challenges was keeping the demo realistic while still being simple enough to run locally.

We had to balance several logistics constraints:

vehicles should be almost full, but not overloaded,
each stop should consult as few pallets as possible,
barrels and heavy returns should stay at the base of the pallet,
delicate products should not have heavy items placed on top,
orders should only be split between pallets when necessary,
the first stop should leave enough margin for returns.

Another challenge was integrating a real PDF delivery note without building a complex OCR system. For this prototype, the PDF is used as the official document, while the structured data is modeled manually from the document to simulate how a real extraction layer could work in the future.

Accomplishments that we're proud of

We are proud that the prototype does not only show a visual dashboard, but also explains the logistics reasoning behind the loading plan.

The application can show where a delivery note is located, which pallet it belongs to, why heavy products are positioned in a specific way, and what the driver or warehouse team should take into account.

We also implemented a safe voice integration approach: the demo can run without consuming external credits, while the real application can be connected to ElevenLabs through a backend proxy.

What we learned

We learned how important it is to connect user experience with operational logic. A logistics app is not useful if it only displays data; it needs to explain decisions clearly to the people using it.

We also learned that separating demo mode from real API usage is essential when working with paid external services. This avoids accidental credit consumption and keeps secrets protected.