HDX Logo

Experiment Engine

Observe Nature

A Portable AI Interface for Science

Experiment Engine

Observe science in its natural state

A Portable AI Interface for Wet Lab Science, Experiment Engine observes an experiment in real time, at the lab bench. EE records every step taken in-situ and assists the human researcher.

Lab setup with camera, sensor box, equipment with bounding boxes, Lab Logs UI, and Sensor Box readings

Lab Logs

Lab Logs are generated that document the end-to-end scientific process. Every action, interaction, and measurement is captured with precise timestamps.

Sensor Data

Improving Lab Documentation and Experimental Reproducibility, and bridging the gap between AI and the physical lab space.

How it works

Experiment Engine runs in your browser with computer vision and AI, using your own camera and, optionally, a Sensor Box.

Use any camera your computer supports (overhead view recommended). You do not need our Setup Kit to get started.

After providing basic context, the system detects relevant objects and continuously logs (predicts) actions, interactions, and deviations in real time.

After completion, users maintain a private or public record of the experiment, including event logs, sensor data, context, files, and video.

Lab bench setup with camera on tripod, scale, pipette, and equipment
Computer vision interface showing detected lab objects with bounding boxes and confidence percentages
EE Sensor Box hardware device
SENSOR DATA
Temp21.1 C
Pressure1013.3 hPa
Humidity22.9%
VOC212
Exp: exp_17719637...

Sensor readings are embedded inside each individual event log

Shop

Optional hardware for the lab bench. You can run Experiment Engine with your own camera and skip the kit entirely.

macOS ยท Free & open source

Sensor Box OS

A dedicated Mac desktop app for the EE Sensor Box โ€” use it on its own, without running Experiment Engine. Plug in your Sensor Box, start a live timed session, and capture every environmental reading (temperature, humidity, pressure, VOC). When you're done, export the session as CSV or JSON.

For now, run the app locally from source on GitHub (Node.js and Rust required). A signed Mac installer will be available later.

Quick start (from the repo README):

git clone https://github.com/avwtr/SensorBox.git
cd SensorBox
npm install
npm run tauri dev

Free and open source โ€” view source, contribute, or fork