What AI believes
Safe-to-trust check
Memory, RAG, policy, and workflow state before they become action basis.
TraceScript → TrustScript → domain surfaces.
TraceScript is the substrate language and runtime — operators, receipts, replay, and invocation history. TrustScript is the security-facing layer on that core. SmartPeers and CSD/DENSITY extend the same receipt-backed runtime to other domains. TrustScript hub → · Buyer whitepaper →
TraceScript
Substrate language + runtime inside the Coordination Operating System.
TraceScript is not a policy wrapper around LLMs. It is a language for governed substrate mutation: every incoming signal is evaluated, every important transition emits a receipt, and replay contracts bind decisions to operator chains and artifact lineage.
- GovernedSignal intake — synchronous evaluation in the API path
- Operator registry — versioned, hash-bound execution primitives
- Receipt ledger — tamper-evident proof with artifact bindings
- Kernel replay — deterministic reconstruction of governance decisions
Runtime architecture → · Technical whitepaper → · Language foundation →
TrustScript
Security and proof layer — first commercial security surface on TraceScript.
TrustScript governs whether signals, memories, policies, workflow states, commitments, and agent actions may become trusted enterprise state. Six product surfaces span action firewall, memory/RAG integrity, policy corpus, workflow truth, continuity verification, and evidence export. Full TrustScript hub →
What AI says
Safe-to-say check
Outbound communications before they create reliance or obligation.
What AI does
Safe-to-act check
Tool release, external writes, merges, and deployments with proof.
First wedge:
Start where enterprise risk is highest — external agent actions and durable memory / retrieval basis. Both surfaces emit receipts and bind to operator chains for deterministic replay. Wedge brief →
Six TrustScript surfaces
All application whitepapers
SmartPeers
Governed peer review and canonicalization.
Enforce claim readiness, section locks, reviewer trust, and paper finalization. Stop premature promotion from draft to canonical.
- Claim readiness and evidence-backed assertions
- Section locks and reviewer trust scores
- Paper finalization readiness gates
- Product dashboard for review posture
Companion application whitepapers for review readiness and canonicalization are planned. Try SmartPeers product-depth routes →
CSD / DENSITY
Governed coordination-state modeling.
Estimate group coordination state, compute pressure fields, and plan interventions — with receipts for every state transition.
- Coordination state estimation
- Pressure field computation
- Intervention planning and outcome capture
- Group dashboard and observability
Companion application whitepapers for coordination-state governance are planned. Try CSD product-depth routes →
Application whitepapers
Deep-dive documents for TrustScript surfaces and TrustScript-governed branches — read online or download Markdown and PDF.
Application whitepapers
Agent Action Firewall, Substrate Integrity Monitor, Policy Corpus Integrity, Workflow Truth and Commitment Ledger, Disclosure Firewall, Code Medium, and Constitutional Agents — public whitepapers plus technical companions.
View applications →