SILENCE.OBJECTS — Monorepo

Open-core framework for structural behavioral analysis — patterns, sequences, rhythms, attention — designed for privacy, neurodiversity, and EU-grade compliance (GDPR + EU AI Act, limited-risk).

Products

Product	Description	Repo
PatternLens B2C	Personal behavioral cockpit — neurodiversity-aware PWA	apps/patternlens
PatternsLab B2B	Team & research workspace for pattern analysis at group level	apps/patternslab
Portal	Unified entry point & auth layer	apps/portal

No diagnostics. No therapy. No performance scoring. Only behavioral patterns and attention profiles, expressed in neutral, structural language.

Quick Start

# Prerequisites: Node 20+, pnpm 9+
pnpm install

# Run PatternLens B2C
pnpm dev:lens

# Run PatternsLab B2B Portal
pnpm dev

# Build all
pnpm build:all

# Lint + type-check all packages
pnpm lint

Monorepo Structure

SILENCE.OBJECTS-monorepo/
├── apps/
│   ├── patternlens/        # B2C PWA (Next.js 15)
│   ├── patternslab/        # B2B workspace (Next.js 15)
│   ├── portal/             # Shared auth portal
│   └── pattern-lens/       # Legacy / reference
├── packages/
│   ├── @silence/           # Core framework packages
│   ├── core/               # Behavioral engine core
│   ├── events/             # SilenceEventV1 schema (frozen post-MVP)
│   ├── contracts/          # Academic Data Contracts (.yaml)
│   ├── language/           # S11/S12 language enforcement
│   ├── ui/                 # Design system & components
│   ├── ai/                 # AI Foundry client (France Central)
│   ├── legal/              # GDPR/EU AI Act compliance utilities
│   └── validator/          # Schema validation
├── agents/                 # AI agent configurations
├── docs/                   # Architecture, ADRs, compliance
├── scripts/                # Build & automation scripts
├── supabase/               # DB migrations & RLS policies
├── pnpm-workspace.yaml     # Workspace definition
├── turbo.json              # Build pipeline (Turbo)
└── package.json            # Root scripts

Architecture

Core Layers (L0–L6)

Layer	Name	Responsibility
L0	Input Capture	Touch, gesture, voice normalization
L1	Pattern Detection	Behavioral sequence recognition
L2	JITAI Engine	Just-In-Time Adaptive Interventions
L3	Phi Measurement	BEH/PER/BIZ wellbeing metrics
L4	Export	GDPR-compliant data portability
L5	AI Inference	Azure AI Foundry — France Central
L6	Observability	Azure Monitor + silence-command-center

Key Architecture Rules

• packages/core/* NEVER imports from enterprise modules (enforced by ESLint + CI)
• SilenceEventV1 schema is frozen post-MVP — no new fields without ADR
• All AI inference routes through Azure AI Foundry France Central (RODO compliance)
• S11 Language Policy is a compliance requirement, not a style guide

Architecture Decision Records (ADRs)

ADR	Title	Status
ADR-001	Control Plane vs Data Plane separation	ACCEPTED
ADR-003	Open-Core licensing model	ACCEPTED
ADR-004	Behavioural Engine as SSOT	ACCEPTED
ADR-005	S11/S12 Language Policy	ACCEPTED
ADR-211	S11 CI enforcement	ACCEPTED
ADR-238	Monorepo consolidation	ACCEPTED
ADR-240	Control plane architecture	ACCEPTED
ADR-250	Behavioral engine 2026 roadmap	ACCEPTED
ADR-254	Scientific engine design	ACCEPTED
ADR-256	Command center integration	ACCEPTED

Benchmark 2030 Metrics

Metric	Target	Current	Status
Test Coverage	≥80%	—	🟡 Sprint 1
S11 Violations in prod	0	0	✅
Critical CVE	0	—	🟡 Scanning
SBOM per release	✅	✅	✅
Deploy frequency	continuous	ad-hoc	🟡 Sprint 2
MTTR	≤2h	unknown	🟡 Sprint 3
Foundry cost/MAU	<0.50 PLN	unknown	🟡 Sprint 2

Compliance

Standard	Status	Details
EU AI Act	Limited Risk	Art. 13 transparency — docs in docs/COMPLIANCE_AI_LIMITED_RISK.md
GDPR	Compliant	France Central region, RLS active, data portability (art. 20)
S11 Language	Enforced by CI	No clinical terms, no simplification language in any output
SBOM	Auto-generated	anchore/sbom-action on every main push

CI/CD — Benchmark 2030 Quality Gate

Every PR triggers:

1. S11 Language Sentinel — blocks forbidden terms (easy, simple, just, basic)
2. Quality Gate — lint + type-check + coverage ≥80%
3. Trivy Security Scan — zero Critical CVE policy
4. SBOM Generation — spdx-json artifact on main push
5. Release Please — auto CHANGELOG + semantic versioning

Development

# Add a package to a specific workspace
pnpm add <pkg> --filter @silence/core

# Run tests in a specific app
pnpm --filter patternlens test

# Type-check everything
pnpm -r exec tsc --noEmit

# Generate SBOM locally
npx @anchore/sbom .

License

Open-core: packages/* — MIT License
Enterprise modules: ee/* — Proprietary, Patternslab-ecosystem © 2026

See LICENSE for details.