Use cases
Home · factory floor
The same tiers, from a front door to a production line.
Two places Furcate is designed for, told step by step: a home camera system that keeps its footage at home and keeps working with the internet down, and a factory floor where robots and machines learn, are watched over and keep their safety local.
01
Home security
Cameras that report events, not footage.
The house keeps its own recordings, and nobody calls in to a camera.
Every camera detects on its own device and passes up only what matters. The home gateway node checks again, keeps the recordings and tells the owner in a sentence. With the internet gone, nothing a camera needs goes with it.
Use case · home security cameras
- 01
Detection on the camera
Each camera leaf node runs its detector on the device. Video stays where it was recorded; only the event and a small crop move up.
- 02
A second look at the gateway
The home gateway node re-checks the event with a larger model and keeps the recordings, in the house.
- 03
A summary, not a stream
The owner gets a short alert: what was seen, where, when and how sure. Footage stays at home unless the owner asks for it.
- 04
Nobody calls in
Nothing connects into a camera unless the owner has authorised it. The cameras dial out to the gateway; nothing dials in to them.
- 05
Internet down, still watching
With the internet down, detection, re-checks and recording carry on. Alerts reach the owner on the home network, and the rest go out when the link returns.
02
The factory floor
Robots and machines, trained, watched over and kept safe under the same tiers.
Every cell learns from its own work, and no cell gives up its data to do it.
Robot arms, CNC machines, conveyors, process equipment and AGVs sit on the same ladder: leaf nodes at each machine, a gateway node per cell, edge AI on the line and the largest models above. Training, monitoring, predictive maintenance and quality inspection all run where the data is, and safety never leaves the leaf.
Use case · robots and machines on a factory floor
- 01
Skills learned in simulation first
A new robot skill is trained and tested in simulation before any cell runs it, so the first real attempt is not the first attempt.
- 02
Every cell learns from its own work
Robots and machines fine-tune on their own cell's experience: grips, cycle times, the parts and the material in front of them.
- 03
Monitoring and maintenance at the machine
Leaf nodes on CNC machines, conveyors, process equipment and AGVs watch vibration, temperature and current. The gateway forecasts wear, so maintenance comes before the failure.
- 04
Inspection at the line
The edge AI node inspects parts as they pass and sends up only the ones it is unsure of.
- 05
Updates up, better models down
Cells send model updates, never their data. Improved models come back signed, through the registry, to every cell.
- 06
Safety stays local
Interlocks, emergency stops and safe states live on the leaf beside each machine. No model update can change them.