Private cloud
Your machines, run as a cloud
A private cloud you own outright.
Furcate turns machines an operator owns into a cloud: sites built from source and served, tenants admitted with an exit, storage, databases and object storage, and models served on accelerators the operator controls. Praecise Furcate adds tenancy, rollout, consolidation and operation across sites. Praecise Cloud runs the same system on Praecise's own machines.
01
The cloud, built in
Services a cloud needs, as properties of the operating system.
No separate control plane to buy, run or trust.
Each service runs on the machines it describes and answers to the operator's policy and keys.
- 01
Publishing
Sites and services built from source, placed on machines you own, served with releases named by the hash of the bytes; rollback is a pointer move.
- 02
Tenants
Workloads a tenant rents on a machine the operator owns: size, term, boundary, and how each would leave.
- 03
Storage and data
Disks, databases, object storage and backups, sealed to the machines that hold them.
- 04
Inference
Models served on an OpenAI-compatible endpoint, one gateway for every local model, loaded on demand and asked of the fleet when a machine cannot answer.
- 05
Placement
Workloads spread across node, chassis, rack, UPS and feed.
- 06
Consolidation
Work packed onto fewer machines and the rest put to sleep, with every saving costed against the grid.
02
Three ways to run it
The same system, wherever it is placed.
Own it, place it at the edge, or let Praecise run it.
Moving between them never changes who holds the keys.
| On your machines | Hybrid with the edge | On Praecise Cloud | |
|---|---|---|---|
| Machines | yours | yours, at many sites | Praecise's, on Praecise Metal |
| Keys | yours | yours | yours |
| Operated by | you, supported by Praecise | you and Praecise | Praecise |
| Fits | a data room, a lab, a private AI cluster | buildings, plants, cities | teams with no hardware to hold |