Answers about Arche

Arche Vision is a browser-based tool that uses computer vision and AI to help document wet-lab work in real time. A camera watches your bench (live feed or uploaded video). You describe the experiment, calibrate which objects are on the bench, then run the session. The system produces timestamped Lab Logs, high-frequency interaction data (Raw Data), optional environmental readings from a Sensor Box, and an end-of-experiment summary you can keep private or share.

No. Arche Vision works with any camera your computer can use (USB webcam, laptop camera, or a tripod-mounted camera above the bench). A bird's-eye view with the full workspace in frame works best. Our Setup Kit (camera, tripod, and accessories) is optional convenience, not a requirement.

The Sensor Box is optional. It records temperature, humidity, pressure, and VOC near your bench and attaches those readings to event logs during live experiments (Chrome or Edge). You can purchase a pre-built Sensor Box, or build your own and run our open-source Sensor Box software. The Mac desktop app is free on GitHub (SensorBox repo), and you can extend or fork it. Arche Vision runs fully without a Sensor Box.

Sign in at archescience.ai, open My Experiments, and start a new experiment. Enter a name, objective, and optional environmental notes; you can link an Arche Lab and choose public or private visibility. Calibrate the bench with AI object detection (live camera or recorded video), adjust labels and bounding boxes, then confirm. Press Start for a live session, or process an uploaded recording. When finished, Conclude, classify the outcome, and review the summary page (video, logs, AI-generated protocol and notes, exports). For live runs, keep the Arche Vision tab open on your computer.

Lab Logs are readable, AI-generated descriptions of what the system infers from the video, roughly on a ~20 second cadence, with some processing delay. Multiple logs can appear from one window if several actions are detected. Raw Data is a separate, higher-volume stream: interaction events fired when a calibrated region of interest (ROI) on the bench is engaged. Raw Data does not automatically create a Lab Log; the two streams complement each other but are not one-to-one.

Arche Labs (archescience.ai) is where you create a lab profile, public or private, to organize members, files, and your research presence. Arche Vision is where you run and record bench experiments. When you create an experiment, you can connect it to an existing Arche Lab so it is associated with that group. If you mark an experiment public, it can be displayed on your lab page for others to discover. Lab members with access can work within the same lab context; private experiments stay limited to you and your lab’s permissions.

The product is in active beta. Treat Lab Logs, summaries, and AI-generated documents as assistive documentation, not a definitive legal or regulatory record. Review outputs, flag lab logs you believe are wrong, add context notes for major changes or issues, and keep your own primary records and backups as you normally would.