Runtime
Local model execution
The internal system can operate with an approved local model and adapt configuration to available hardware.
006ISACFirst product on GCF
A persistent AI system where identity is bigger than the model.
ISAC is designed to remember, use tools and work across cognitive engines while GCF governs the authority around those capabilities.
The model is replaceable.
A model can be upgraded, swapped or selected to match the available hardware. ISAC's wider identity, memory, tools and continuity stay with the system.
- Persistent identity across model changes.
- Governed memory rather than unmanaged context alone.
- Hardware-aware cognition that adapts to the resources available.
- Scoped tools controlled through the GCF boundary.
- Local-first baseline with external/custom providers planned later.
$ isac status
runtime ................. online
model .................. local cognitive engine
memory ................. governed · persistent
tools .................. scoped
credentials ............ isolated
authority policy ....... active
// ISAC's job is to make the architecture useful in practice.
WorkingWhat exists
Working internally, not commercially launched.
This distinction should stay explicit throughout the website.
Governance
Authority controls
Policy and governance checks are part of the system before sensitive actions are allowed to proceed.
Next
External validation
After foundation hardening, the project moves toward controlled evaluation, security testing and later Early Access.
↗Open evidence
Read how the architecture actually works.
A full technical breakdown of the runtime, memory, governance boundary and scoring function. Published so engineers, security researchers and investors can verify the design claims.
Open technical evidence ↗// PDF · 38 pages · last revised 2026-09-22 · no email gate