Private deployment
Run Flash where your security team allows.
Flash works against local model backends, so the whole loop — analysis and inference — can stay inside your environment. Cloud is an option you turn on, never a requirement.
Private cloud
Deploy inside your own VPC, alongside the infrastructure your team already governs.
On-prem
Run on hardware you control, behind your own firewall and identity, with nothing phoning home.
Air-gapped
No outbound connectivity required at all — the deterministic analysis and a local model run fully offline.
How the boundary works
Schematics · firmware · logs · datasheets
Deterministic context engine
Local model
Cloud model
off by default
Air-gapped mode: no outbound connectivity at all — not even the optional line above.
Bring your own model
Flash is provider-agnostic. Point it at a local model runtime or any OpenAI-compatible endpoint you host, and switch models without changing how you work. Use a cloud provider only on the tasks where you decide it's appropriate.
And it adapts to your stack — your SDKs, RTOS, internal BSPs and libraries, coding standards, and validation workflows — rather than forcing a generic one.