Local-first edge software for synchronized, event-tagged camera and sensor data capture from any vendor: offline, with your raw data and models kept in-house

What it is. Banalytics is local-first edge software that captures synchronized, event-tagged data from cameras and sensors of any vendor (ONVIF and RTSP cameras, OPC UA, Modbus RTU/TCP, MQTT v3/v5, ZeroMQ, USB, I2C). It runs offline on your own hardware. Raw data, models and algorithms stay with you.

Why it matters. R&D and process teams often lose weeks stitching together camera streams and sensor signals that don’t share a clock or a format. The result is unlabeled, misaligned data that is hard to use for process analysis or model training. Banalytics aligns the streams in time, tags events, keeps the history, and exports it to MQTT, local files or Google Cloud Storage.

How it works. For industrial sensing, your own algorithm runs inside a locked module interface. Only structured outputs, health status and timestamps leave the module, and your code stays private. Dashboards, an Event History and MQTT output to SCADA are included. For AI data teams, Banalytics provides the full pipeline from capture to labelled datasets. Retention settings, access control and an audit log of views and exports are built in.

Proof to date. The software is released and in production use. We also deliver engineering integration work for industrial and sensing projects.

Typical use cases. Synchronizing camera video with PLC and sensor signals on a production line to find the cause of quality events. Recording labelled multi-sensor datasets from test rigs and field equipment for model training. Keeping a local, auditable history of events from existing cameras and sensors without sending data to the cloud.

Next steps. We start with a free technical demo, then a free proof of concept on your recorded or mock data, then a funded on-site pilot. We are open to pilot partners, co-development and integration partners, and supply or licence agreements.