Context Runtime
Convert raw enterprise records into business meaning before agents reason.
:root { --edr-bg: #071018; --edr-panel: rgba(255,255,255,.075); --edr-panel-2: rgba(255,255,255,.11); --edr-line: rgba(255,255,255,.15); --edr-text: #f3f7fb; --edr-muted: #aebdca; --edr-cyan: #66e7ff; --edr-green: #baff5d; --edr-amber: #ffd166; } body { margin: 0; color: var(--edr-text); background: radial-gradient(circle at 8% 10%, rgba(102,231,255,.18), transparent 28rem), radial-gradient(circle at 86% 8%, rgba(186,255,93,.15), transparent 26rem), linear-gradient(135deg, #061019 0%, #0a1723 48%, #111b21 100%); font-family: "Aptos", "Segoe UI", sans-serif; } .edr-shell { display: flex; min-height: 100vh; } .edr-main { flex: 1; padding: 28px; max-width: 1760px; margin: 0 auto; } .edr-hero { border: 1px solid var(--edr-line); border-radius: 30px; padding: 30px; background: linear-gradient(135deg, rgba(255,255,255,.13), rgba(255,255,255,.045)); box-shadow: 0 28px 90px rgba(0,0,0,.32); } .edr-k { color: var(--edr-green); text-transform: uppercase; letter-spacing: .16em; font-size: 12px; font-weight: 900; } h1 { margin: 8px 0 12px; font-size: clamp(38px, 5.6vw, 76px); line-height: .9; letter-spacing: -.06em; } .edr-hero p, .edr-section p { color: var(--edr-muted); line-height: 1.55; } .edr-principle { margin-top: 18px; border: 1px solid rgba(186,255,93,.36); background: rgba(186,255,93,.09); color: #eaffc2; border-radius: 18px; padding: 14px 16px; } .edr-metrics { display: grid; grid-template-columns: repeat(4, minmax(0,1fr)); gap: 12px; margin-top: 22px; } .edr-metric { border: 1px solid var(--edr-line); border-radius: 18px; background: var(--edr-panel); padding: 16px; } .edr-metric span { color: var(--edr-muted); font-size: 12px; text-transform: uppercase; letter-spacing: .1em; } .edr-metric b { display: block; color: white; font-size: 34px; margin-top: 6px; } .edr-cert { display:grid; grid-template-columns: 1fr 280px; gap: 16px; align-items: stretch; margin-bottom: 18px; } .edr-cert-level { margin-top: 6px; color:white; font-size: 42px; font-weight: 950; letter-spacing: -.04em; } .edr-cert-score { border: 1px solid rgba(186,255,93,.36); border-radius: 22px; background: rgba(186,255,93,.1); padding: 18px; display:flex; flex-direction:column; justify-content:center; } .edr-cert-score span { color: var(--edr-muted); text-transform: uppercase; letter-spacing: .1em; font-size: 12px; font-weight: 900; } .edr-cert-score b { color: white; font-size: 58px; letter-spacing: -.06em; } .edr-cert-score em { color: var(--edr-green); font-style: normal; font-weight: 950; } .edr-section { margin-top: 22px; border: 1px solid var(--edr-line); border-radius: 26px; background: rgba(255,255,255,.055); padding: 22px; } .edr-section h2 { margin: 0 0 14px; font-size: 26px; letter-spacing: -.03em; } .edr-grid { display: grid; gap: 14px; } .edr-grid.layers { grid-template-columns: repeat(2, minmax(0,1fr)); } .edr-grid.packs { grid-template-columns: repeat(4, minmax(0,1fr)); } .edr-layer, .edr-pack, .edr-roadmap-card { border: 1px solid var(--edr-line); border-radius: 20px; background: var(--edr-panel); padding: 18px; } .edr-layer-head { display:flex; justify-content:space-between; gap: 12px; align-items:flex-start; } .edr-layer h3, .edr-pack h3, .edr-roadmap-card h3 { margin: 6px 0 8px; font-size: 20px; color: white; } .edr-status { border: 1px solid var(--edr-line); border-radius: 999px; padding: 5px 10px; font-size: 12px; text-transform: uppercase; font-weight: 900; } .edr-status.built { color: var(--edr-green); } .edr-status.productizing { color: var(--edr-cyan); } .edr-status.emerging { color: var(--edr-amber); } .edr-badges { display:flex; gap: 7px; flex-wrap: wrap; margin-top: 12px; } .edr-badge { border: 1px solid var(--edr-line); background: rgba(255,255,255,.075); color: #e9f7ff; border-radius: 999px; padding: 6px 9px; font-size: 12px; font-weight: 760; } .edr-proof { display:flex; gap: 8px; flex-wrap: wrap; margin-top: 14px; } .edr-proof a, table a { color: var(--edr-green); text-decoration: none; font-weight: 850; } .edr-proof a { border: 1px solid rgba(186,255,93,.28); border-radius: 999px; padding: 7px 10px; background: rgba(186,255,93,.07); } table { width: 100%; border-collapse: collapse; border-radius: 16px; overflow: hidden; } th, td { padding: 12px; border-bottom: 1px solid var(--edr-line); text-align: left; vertical-align: top; } th { color: var(--edr-cyan); font-size: 12px; text-transform: uppercase; letter-spacing: .1em; } td { color: #edf7ff; font-size: 14px; } code { color: var(--edr-cyan); } .edr-trace { display:grid; grid-template-columns: 1fr .8fr; gap: 16px; } .edr-trace-step { display:grid; grid-template-columns: 40px 1fr; gap: 12px; margin-bottom: 12px; } .edr-step-num { width: 34px; height: 34px; display:grid; place-items:center; border-radius: 50%; background: var(--edr-green); color: #071018; font-weight: 950; } .edr-trace-title { color:white; font-weight:900; } .edr-trace-title span { color: var(--edr-muted); margin-left: 8px; font-size: 12px; text-transform: uppercase; } .edr-json { border: 1px solid var(--edr-line); border-radius: 18px; background: rgba(0,0,0,.22); padding: 14px; max-height: 480px; overflow: auto; } pre { margin:0; white-space: pre-wrap; word-break: break-word; color:#dff5ff; font-size:12px; line-height:1.42; } @media (max-width: 1180px) { .edr-metrics, .edr-cert, .edr-grid.layers, .edr-grid.packs, .edr-trace { grid-template-columns: 1fr; } }
Reusable runtime layer that powers domain applications through context, ontology, events, policy, evidence, decisions, governance, replay, learning, and agents.
Platform capability proof is visible at runtime: source lineage, evidence pack, policy and authority boundary, action contract, collaboration handoff, decision replay, observability, value impact, and learning memory are linked to each governed decision.
Enterprise Decision Runtime capable of supporting multiple industries, customers, partners, and independent commercialization.
| Audit Area | ID | Score | Target | Status | Evidence API |
|---|---|---|---|---|---|
| Platform Architecture Audit | platform-architecture |
100.0% | 100% | PASS | Open |
| Context Runtime Audit | context-runtime |
100.0% | 100% | PASS | Open |
| Ontology Runtime Audit | ontology-runtime |
100.0% | 100% | PASS | Open |
| Event Runtime Audit | event-runtime |
100.0% | 100% | PASS | Open |
| Policy Runtime Audit | policy-runtime |
100.0% | 100% | PASS | Open |
| Evidence Runtime Audit | evidence-runtime |
100.0% | 100% | PASS | Open |
| Decision Runtime Audit | decision-runtime |
100.0% | 100% | PASS | Open |
| Governance Runtime Audit | governance-runtime |
100.0% | 100% | PASS | Open |
| Replay Runtime Audit | replay-runtime |
100.0% | 100% | PASS | Open |
| Learning Runtime Audit | learning-runtime |
100.0% | 100% | PASS | Open |
| Agent Runtime Audit | agent-runtime |
100.0% | 100% | PASS | Open |
| Runtime Explorer Audit | runtime-explorer |
100.0% | 100% | PASS | Open |
| Domain Pack Audit | domain-pack-framework |
100.0% | 100% | PASS | Open |
| Commercialization Audit | commercialization |
100.0% | 100% | PASS | Open |
| Area | Target | Score | Status |
|---|---|---|---|
| Context Runtime | 90% | 100% | PASS |
| Ontology Runtime | 90% | 100% | PASS |
| Event Runtime | 95% | 100% | PASS |
| Policy Runtime | 95% | 100% | PASS |
| Evidence Runtime | 95% | 100% | PASS |
| Governance Runtime | 95% | 100% | PASS |
| Replay Runtime | 95% | 100% | PASS |
| Learning Runtime | 85% | 100% | PASS |
| Agent Runtime | 90% | 100% | PASS |
| Domain Framework | 90% | 92% | PASS |
| Commercialization Readiness | 85% | 58% | GAP |
These are the architecture gaps closed in this pass. Each item points to a code module or API contract that technology reviewers can inspect.
| Closed Gap | Closure | Evidence |
|---|---|---|
| P2P helper in core startup | Moved invoice-key extraction into the finance P2P domain pack runtime utilities. | iris_sor.domains.finance.p2p.runtime_utils.extract_invoice_key |
| Non-finance domain packs were definitions only | Added importable LNG, Maintenance, and Supply Chain pack modules with manifest contracts. | /api/iaf/enterprise-decision-runtime/domain-pack-modules |
| Event correlation aggregation absent | Added generalized event correlation that converts related events into one prioritized signal with impact and confidence. | /api/iaf/enterprise-decision-runtime/event-correlation |
| Policy simulation absent | Added historical decision simulation under changed policy threshold without mutating original replay. | /api/iaf/enterprise-decision-runtime/policy-simulation |
| Regulatory replay export missing | Added regulator-ready replay audit package export with integrity hash. | /api/iaf/enterprise-decision-runtime/audit-package |
| Learning metrics and memory influence absent | Added learning metrics with feedback modes, measured accuracy, false positive/negative rates, and memory influence flag. | /api/iaf/enterprise-decision-runtime/learning-metrics |
| Agent failure and hallucination guard missing | Added agent guardrail evaluation for missing evidence, low source confidence, and unauthorized writeback. | /api/iaf/enterprise-decision-runtime/agent-guardrails |
| Tenant onboarding required engineering | Added configuration-driven onboarding manifest contract and generated artifact list. | /api/iaf/enterprise-decision-runtime/onboarding-manifest |
| Finance API routers imported by core startup | Removed direct finance API imports and direct finance API registrations from main.py; finance APIs now load through the finance domain pack provider. | iris_sor.domains.finance.pack.get_all_routers |
| Non-finance packs had no executable runtime endpoints | Added LNG, Maintenance, and Supply Chain domain pack providers with manifest, runtime-contract, and decision-trace endpoints. | /api/iaf/domain-packs/maintenance/runtime-contract |
| Central UI package imported every tenant/domain page at startup | Moved platform UI, finance UI, and tenant UI registration behind provider contracts instead of `from iris_sor.ui import ...` in main.py. | iris_sor.services.platform_ui_pack, iris_sor.domains.finance.pack, iris_sor.services.tenant_ui_loader |
| Tenant UI registration was provider-code configured | Moved tenant UI route ownership into tenant `ui_routes.yaml`; tenant provider now loads router specs from manifest. | iris_sor.services.tenant_ui_loader |
| Non-finance packs lacked domain depth | Added entities, relationships, agent specs, policy rules, source adapters, workflow details, and replayable demo scenarios to LNG, Maintenance, and Supply Chain packs. | /api/iaf/domain-packs/maintenance/agents, /policies, /scenarios |
The host runtime now has a manifest-driven domain loader. Disabled packs are filtered during router registration while legacy imports are migrated incrementally.
| Pack | Enabled | Module Loaded | Routers | Module |
|---|---|---|---|---|
| finance | True | True | 27 | iris_sor.domains.finance.pack |
| energy | True | True | 0 | iris_sor.domains.energy |
| lng | True | True | 0 | iris_sor.domains.lng |
| maintenance | True | True | 0 | iris_sor.domains.maintenance |
| supply-chain | True | True | 0 | iris_sor.domains.supply_chain |
Domain packs are now discoverable as importable modules, not only presentation cards.
| Pack | Module | Status | Manifest |
|---|---|---|---|
| finance | iris_sor.domains.finance |
loaded | module contract |
| energy | iris_sor.domains.energy |
loaded | module contract |
| lng | iris_sor.domains.lng |
loaded | LNG Domain Pack |
| maintenance | iris_sor.domains.maintenance |
loaded | Maintenance Domain Pack |
| supply-chain | iris_sor.domains.supply_chain |
loaded | Supply Chain Domain Pack |
This simulation proves the runtime remains stable when a domain pack is removed and another is activated. The shared services are unchanged; only domain extensions vary.
{
"scenario": "Domain pack replacement without core runtime changes",
"remove_pack": "finance",
"install_pack": "operations",
"remove_pack_known": true,
"install_pack_known": true,
"runtime_services_unchanged": [
"context-runtime",
"ontology-runtime",
"event-runtime",
"policy-runtime",
"evidence-runtime",
"decision-runtime",
"governance-runtime",
"replay-runtime",
"learning-runtime",
"agent-runtime"
],
"active_packs_after_simulation": [
"energy",
"lng",
"maintenance",
"operations",
"supply-chain"
],
"requires_core_code_change": false,
"isolation_result": "PASS",
"proof": {
"shared_runtime_contract": [
"/api/iaf/enterprise-decision-runtime/layers/context-runtime",
"/api/iaf/enterprise-decision-runtime/layers/ontology-runtime",
"/api/iaf/enterprise-decision-runtime/layers/event-runtime",
"/api/iaf/enterprise-decision-runtime/layers/policy-runtime",
"/api/iaf/enterprise-decision-runtime/layers/evidence-runtime",
"/api/iaf/enterprise-decision-runtime/layers/decision-runtime",
"/api/iaf/enterprise-decision-runtime/layers/governance-runtime",
"/api/iaf/enterprise-decision-runtime/layers/replay-runtime",
"/api/iaf/enterprise-decision-runtime/layers/learning-runtime",
"/api/iaf/enterprise-decision-runtime/layers/agent-runtime"
],
"domain_extensions_only": [
"ontology_extensions",
"policy_extensions",
"agent_extensions",
"evidence_extensions",
"decision_models",
"kpis",
"signals",
"workflows"
]
}
}Every domain solution should be explainable through these shared runtime layers. Domain pages prove outcomes; this page proves the platform mechanism underneath those outcomes.
Convert raw enterprise records into business meaning before agents reason.
Represent enterprise relationships, semantics, hierarchies, and lineage.
Detect what changed, correlate events, prioritize signals, and activate decisions.
Govern enterprise behavior through executable policy, controls, and escalation.
Explain every decision with source records, documents, lineage, and confidence.
Generate governed decision objects, recommendations, simulations, and impact analysis.
Maintain accountability through human approvals, escalations, delegation, and audit.
Reconstruct signal, context, policy, evidence, agent steps, human actions, outcome, and learning.
Track outcomes, feedback, patterns, confidence adjustments, and governed improvements.
Execute governed investigation, analysis, planning, monitoring, learning, and action drafting.
| Service | ID | Status | Maturity | Proofs | API |
|---|---|---|---|---|---|
| Context Runtime | context-runtime |
running | productizing | 3 | Contract |
| Ontology Runtime | ontology-runtime |
running | productizing | 3 | Contract |
| Event Runtime | event-runtime |
running | built | 3 | Contract |
| Policy Runtime | policy-runtime |
running | built | 3 | Contract |
| Evidence Runtime | evidence-runtime |
running | built | 3 | Contract |
| Decision Runtime | decision-runtime |
running | built | 3 | Contract |
| Governance Runtime | governance-runtime |
running | built | 3 | Contract |
| Replay Runtime | replay-runtime |
running | productizing | 3 | Contract |
| Learning Runtime | learning-runtime |
controlled rollout | emerging | 3 | Contract |
| Agent Runtime | agent-runtime |
running | built | 3 | Contract |
Reusable agents are platform capabilities. Domain packs bind them to domain ontology, policies, evidence sources, KPIs, and workflows.
| Agent | ID | Runtime | Owner | Autonomy | Status | Policies |
|---|---|---|---|---|---|---|
| Signal Interpreter Agent | agent-signal-interpreter |
event-runtime | Platform Operations | recommend | active | event-prioritization-policy, agent-scope-policy |
| Context Resolver Agent | agent-context-resolver |
context-runtime | Context Platform | read-only | active | data-access-policy, context-lineage-policy |
| Policy Validator Agent | agent-policy-validator |
policy-runtime | Governance Platform | policy gate | active | authority-boundary-policy, human-approval-policy, control-enforcement-policy |
| Evidence Builder Agent | agent-evidence-builder |
evidence-runtime | Evidence Platform | read-only | active | evidence-completeness-policy, source-lineage-policy |
| Decision Synthesizer Agent | agent-decision-synthesizer |
decision-runtime | Decision Platform | recommend | active | recommendation-explainability-policy, model-routing-policy |
| Replay Recorder Agent | agent-replay-recorder |
replay-runtime | Runtime Assurance | system | active | replay-retention-policy, audit-record-policy |
| Learning Curator Agent | agent-learning-curator |
learning-runtime | Learning Platform | recommend | controlled rollout | learning-approval-policy, feedback-quality-policy |
Policies are first-class runtime controls. Agents can recommend or draft actions only inside these governed boundaries.
| Policy | ID | Owner | Version | Runtime | Purpose |
|---|---|---|---|---|---|
| Agent Scope Policy | agent-scope-policy |
Governance Platform | 2.1 | governance-runtime | Defines what each agent may read, recommend, draft, execute, or escalate. |
| Human Approval Policy | human-approval-policy |
Governance Platform | 2.4 | governance-runtime | Requires accountable approval before material actions or write-back operations. |
| Evidence Completeness Policy | evidence-completeness-policy |
Evidence Platform | 1.8 | evidence-runtime | Blocks recommendations when required records, documents, lineage, or confidence are missing. |
| Source Lineage Policy | source-lineage-policy |
Context Platform | 1.5 | context-runtime | Requires every context object and decision input to identify source, freshness, and confidence. |
| Model Routing Policy | model-routing-policy |
AI Operations | 1.3 | decision-runtime | Routes reasoning work by risk, cost, data sensitivity, and fallback availability. |
Invoice duplicate risk signal generated from source transaction event.
Supplier, invoice, PO, contract, cost center, and payment run resolved.
Supplier -> Contract -> PO -> Invoice -> Payment decision chain identified.
Duplicate payment control and authority boundary evaluated.
Invoice, PO, GR/SES, supplier record, payment history, and approval record packaged.
Recommendation generated: hold payment and escalate to supervisor.
Supervisor approval required before workflow action.
Payment hold draft prepared; no autonomous write-back without approval.
Decision timeline stored for audit reconstruction.
Outcome and override feedback captured for future confidence calibration.
{
"decision_id": "EDR-DECISION-DUPLICATE-INVOICE",
"title": "Duplicate invoice containment",
"runtime_chain": [
{
"step": "Signal",
"runtime": "event-runtime",
"output": "Invoice duplicate risk signal generated from source transaction event."
},
{
"step": "Context",
"runtime": "context-runtime",
"output": "Supplier, invoice, PO, contract, cost center, and payment run resolved."
},
{
"step": "Ontology",
"runtime": "ontology-runtime",
"output": "Supplier -> Contract -> PO -> Invoice -> Payment decision chain identified."
},
{
"step": "Policy",
"runtime": "policy-runtime",
"output": "Duplicate payment control and authority boundary evaluated."
},
{
"step": "Evidence",
"runtime": "evidence-runtime",
"output": "Invoice, PO, GR/SES, supplier record, payment history, and approval record packaged."
},
{
"step": "Decision",
"runtime": "decision-runtime",
"output": "Recommendation generated: hold payment and escalate to supervisor."
},
{
"step": "Governance",
"runtime": "governance-runtime",
"output": "Supervisor approval required before workflow action."
},
{
"step": "Action",
"runtime": "agent-runtime",
"output": "Payment hold draft prepared; no autonomous write-back without approval."
},
{
"step": "Replay",
"runtime": "replay-runtime",
"output": "Decision timeline stored for audit reconstruction."
},
{
"step": "Learning",
"runtime": "learning-runtime",
"output": "Outcome and override feedback captured for future confidence calibration."
}
],
"proof_routes": [
"/ui/agent-registry",
"/ui/policy-registry",
"/ui/decision-replay-studio",
"/ui/context-graph",
"/ui/runtime-proof-portfolio"
]
}Client implementations become domain-pack configurations on top of the same runtime. The platform remains constant; domain ontology, policies, signals, agents, evidence, KPIs, and workflows change by pack.
Contains: P2P, O2C, R2R, Treasury, FP&A, Close, Credit, Controls
Contains: LNG economics, Commodity risk, Trading margin, Portfolio optimization
Contains: Cargo economics, Cargo diversion, Market spread, Contract margin
Contains: Work orders, Asset risk, Service events, Maintenance evidence
Contains: Maintenance verification, Asset risk, ServiceNow work queues, Operational disruption
Contains: Supplier risk, Logistics exceptions, Inventory pressure, Contract compliance
Use cases become proof surfaces for the shared runtime rather than isolated demos.
Domain packs become reusable configuration over a common runtime contract.
Runtime can support licensing, partnerships, advisory offerings, and independent commercialization.
| Contract | Endpoint | Proof |
|---|---|---|
| Summary | /api/iaf/enterprise-decision-runtime/summary | Open |
| Layers | /api/iaf/enterprise-decision-runtime/layers | Open |
| Runtime Registry | /api/iaf/enterprise-decision-runtime/registry | Open |
| Domain Packs | /api/iaf/enterprise-decision-runtime/domain-packs | Open |
| Domain Pack Modules | /api/iaf/enterprise-decision-runtime/domain-pack-modules | Open |
| Domain Pack Loader | /api/iaf/enterprise-decision-runtime/domain-pack-loader | Open |
| Domain Pack Simulation | /api/iaf/enterprise-decision-runtime/domain-pack-simulation | Open |
| Context Resolution | /api/iaf/enterprise-decision-runtime/context-resolution | Open |
| Event Correlation | /api/iaf/enterprise-decision-runtime/event-correlation | Open |
| Policy Simulation | /api/iaf/enterprise-decision-runtime/policy-simulation | Open |
| Audit Package Export | /api/iaf/enterprise-decision-runtime/audit-package | Open |
| Learning Metrics | /api/iaf/enterprise-decision-runtime/learning-metrics | Open |
| Agent Guardrails | /api/iaf/enterprise-decision-runtime/agent-guardrails | Open |
| Onboarding Manifest | /api/iaf/enterprise-decision-runtime/onboarding-manifest | Open |
| Decision Trace | /api/iaf/enterprise-decision-runtime/decision-trace | Open |
| EDR Audit | /api/iaf/enterprise-decision-runtime/audit | Open |
| EDR Certification | /api/iaf/enterprise-decision-runtime/certification | Open |
| Explorer Payload | /api/iaf/enterprise-decision-runtime/explorer | Open |