Inspiration
Pharma distributors lose entire shipments to temperature excursions that nobody catches until the truck is unloaded. Paper logs and end-of-route data downloads mean a breach discovered on Friday happened on Tuesday. We built ColdTrace to close that gap: catch the breach in the moment, and hand over a compliance log regulators accept without argument.
What it does
ColdTrace reads from BLE temperature sensors placed inside shipping containers, streams the data through an Android gateway to a cloud backend, and fires an SMS and dashboard alert the second a reading crosses a configured threshold. Every shipment ends with a signed, tamper-evident PDF report showing the full temperature history, breach events, and corrective actions logged by the driver.
How we built it
Sensors push readings over BLE to an Android gateway app written in Kotlin. The gateway buffers data locally and syncs it via MQTT to a Node.js ingestion service backed by TimescaleDB. Alert rules run as a lightweight stream processor on the ingestion path. The manager dashboard is React with WebSocket updates, and reports are generated server-side and cryptographically signed.
Challenges we ran into
Long-haul routes drop connectivity for hours, so offline buffering and idempotent sync had to be bulletproof. Sensor clock drift produced false breach timestamps until we moved timestamping to the gateway. Getting alert latency under 30 seconds on 2G networks took multiple rounds of payload compression and retry tuning.
Accomplishments that we're proud of
Sub-minute breach alerts across a 40-vehicle pilot. Zero duplicate or missing readings after the sync redesign. A report format that cleared a distributor's internal GDP audit on the first submission.
What we learned
In logistics, reliability at the edge beats dashboard polish every time. Most of our engineering hours went into the sync layer nobody sees, and that is exactly what made the product trusted.
What's next for ColdTrace
Predictive breach alerts using route data and ambient weather, plus a driver-facing app that prompts corrective actions in real time.
Built With
kotlin · android · ble · mqtt · node.js · timescaledb · react · websockets · aws