Portable current context
Package selected project observations, source locators and currentness metadata into one deterministic Snapshot IR.
Private product · portable context
RUAP
Portable AI Continuity & Governance for long-running agent work. RUAP turns selected, provenance-aware project context into deterministic cold-start packs that can move between models, agents and sessions without turning remembered context into permission to act.
The problem
AI work should survive a model, chat or tool change.
Long-running projects accumulate decisions, source references, constraints and stale assumptions across many surfaces. RUAP creates a portable snapshot with explicit provenance and freshness boundaries so the next agent can cold-start from evidence instead of reconstructing the project from memory.
Safe context profiles
Create trusted internal packs and external-safe packs with a fail-closed boundary for sensitive material.
Cold-start verification
Verify and validate the pack before transfer, then require the receiving agent to keep provenance, freshness and authority boundaries explicit.
Mental model
Evidence travels. Authority does not.
01Fresh evidencesources · receipts · observations
→
02Snapshotprovenance + freshness
→
03Portable packtrusted / external-safe
→
04New agentcontext, not permission
Useful for
Complex workflows that cannot fit in one chat.
Cross-model handoffs, multi-project operators, long-running technical work, safe contractor context, and ContinuityOS / Control Center integrations.
Boundary
No automatic authority.
A newer RUAP pack does not prove a deployment, runtime effect or permission to execute. Consequential actions still require fresh provider truth and the applicable owner authority.
Standalone + module
Use RUAP alone or as a portability layer.
Standalone: local snapshot, cold-start, safe/trusted bundle and verification workflow.
With ContinuityOS: import/export portable context while checkpoints, proof and replay remain in the continuity layer.
With Control Center: provide portable observations to a decision/orchestration system without granting it new effect authority.
Design-partner delivery
Start with one real fragmented workflow.
A useful first deployment is not a generic memory demo. It is a bounded workflow where project state is spread across several tools or AI sessions and must be reconstructed, verified, packaged and resumed safely.
Scope a RUAP setup