Sovereign Storefront & Platform Evidence Console

Storefront is the critical public revenue, product catalog, autonomous optimization, and conversion telemetry gateway in the Hardonia / AI Automated Systems sovereign software stack.

It presents an ultra-fast, local-first catalog of air-gapped AI tools, autonomous workflow suites, deterministic governance packages, and private GPU compute instances — backed by local SQLite persistence, zero cloud analytics leaks, Bayesian multi-armed bandit learning, offline cryptographic license issuance, and strict browser security defenses.

---

🏛️ System Architecture

Storefront is built as a domain-driven, modular FastAPI application decoupled into clean layers:

storefront/
├── app/
│   ├── core/                  # Core foundations & runtime configuration
│   │   ├── config.py          # Typed Pydantic Settings & stack connective tissue
│   │   ├── database.py        # SQLite connection pool with WAL mode & 15s timeout
│   │   ├── security.py        # HMAC-SHA256 tokens, URL defense, license signing
│   │   └── templates.py       # Jinja2 environment & custom filters
│   │
│   ├── middleware/            # High-performance ASGI middleware stack
│   │   ├── security_headers.py# CSP, X-Frame-Options: DENY, Referrer & Permission policies
│   │   ├── cors_and_limits.py # Strict CORS origin allowlist & 64KB payload bounds
│   │   ├── rate_limiter.py    # Per-IP token bucket rate limiting with honeypot traps
│   │   ├── cache_control.py   # Immutable asset caching & no-store payment controls
│   │   └── request_context.py # Session resolution, request tracking, bot classification
│   │
│   ├── services/              # Autonomous & business domain services
│   │   ├── finetuning_service.py # LoRA/QLoRA VRAM budgeting, dataset token sizing, Unsloth recipes
│   │   ├── benchmark_service.py  # TTFT, TPS throughput, quantization retention & KV cache scaling
│   │   ├── bandit_service.py     # Bayesian Beta-Bernoulli Thompson Sampling bandit
│   │   ├── anomaly_detector.py   # Real-time funnel anomaly & conversion stall detector
│   │   ├── stack_bridge.py       # revenue-os, Hermes Ops Nerve Center, Model Zoo & Compute API bridge
│   │   ├── demand_intelligence.py# Customer inquiry mining & intent gap signals
│   │   ├── license_service.py    # Air-gapped cryptographic software license engine
│   │   ├── product_service.py    # Catalog querying, sorting, enrichment
│   │   └── analytics_service.py  # SQLite events & aggregation
│   │
│   ├── routers/               # Domain-driven API & web controllers
│   │   ├── catalog.py         # Home grid, capability filtering, /p/{slug}, /pricing
│   │   ├── finetuning.py      # /tools/finetuning-optimizer, /api/tools/finetune-estimate, recipe export
│   │   ├── tools.py           # /tools/model-benchmarks, /tools/stack-matrix, /tools/redaction-sandbox, /tools/hardware-sizer
│   │   ├── blueprint.py       # Tailored Sovereign Architecture Blueprint generator (/blueprint/{token})
│   │   ├── commerce.py        # Checkout routing, /buyer locker, /order/success, /download/{slug}
│   │   ├── api_products.py    # Public sanitized JSON product catalog
│   │   ├── api_leads.py       # Lead capture, newsletter subscribe, queued privacy erasure
│   │   ├── api_analytics.py   # Privacy funnel truth ingestion, conversion metrics
│   │   ├── api_flags.py       # Live feature flags, A/B experiments, /api/flags/bandit report
│   │   ├── api_support.py     # AI support assistant proxy & /support-widget.js
│   │   ├── status.py          # /health, /status.json, /proof-score, /api/stack/fleet
│   │   ├── legal.py           # Markdown legal reader with directory traversal protection
│   │   ├── blog.py            # Local AI ops field notes & RSS feed
│   │   ├── seo.py             # robots.txt, sitemap.xml, llms.txt, IndexNow, Google verification
│   │   └── private_ai_ops.py  # Private AI operations landing & synthetic demo suite
│   │
│   └── main.py                # Application entrypoint & middleware assembly
│
├── static/                    # Fast static CSS, JS, SVGs, and landing assets
├── tests/                     # 100% passing test suite (121 contracts)
│   ├── test_finetuning_service.py
│   ├── test_benchmark_service.py
│   ├── test_expanded_tools_routes.py
│   ├── test_threat_model_qa.py
│   ├── test_bandit_service.py
│   ├── test_anomaly_detector.py
│   ├── test_interactive_tools.py
│   ├── test_buyer_locker.py
│   ├── test_storefront.py
│   ├── test_flags.py
│   ├── test_funnel_truth.py
│   ├── test_private_ai_operations_public.py
│   ├── test_observability.py
│   └── test_security_contract.py
│
├── scripts/                   # Operational & developer tooling
│   ├── healthcheck.py         # CLI deployment health validator
│   ├── simulate_stack_bridge.py # Sovereign stack bridge & fleet telemetry probe
│   ├── run_bandit_evaluation.py # Thompson Sampling Monte Carlo audit
│   └── seed_sample_catalog.py # Sample catalog seeder
│
├── systemd/                   # Native Linux systemd service unit
│   └── storefront.service
├── Dockerfile                 # Multi-stage hardened production container
├── docker-compose.yml         # Containerized orchestration
├── run.ps1                    # PowerShell launcher (Windows)
└── run.sh                     # Bash launcher (Linux/macOS)

---

⚡ Autonomous Capabilities & Sovereign Features

1. Bayesian Multi-Armed Bandit (Thompson Sampling)

• Replaces static A/B testing with continuous Bayesian optimization for headlines (hero_variant) and CTAs (cta_variant).
• Samples from $\text{Beta}(1 + \alpha, 1 + \beta)$ posteriors to balance exploration with exploitation.
• Autonomous Winner Promotion: When a variant reaches $\ge 99%$ win probability with sufficient sample size, Storefront automatically pins the winning copy without manual intervention.

2. Real-Time Anomaly & Conversion Stall Detector

• Monitors conversion funnels (landing $\to$ lead_start $\to$ checkout_start $\to$ provider_payment).
• Emits structured operational alerts if checkout abandonment spikes or if 5xx errors surge.

3. Interactive Zero-Cloud Previews & Sizers

• Sentinel Redaction Sandbox (/tools/redaction-sandbox): Live regex/WASM sanitization demo for HIPAA/GDPR clinical and legal notes.
• LLM Hardware Topology Sizer (/tools/hardware-sizer): Calculates exact VRAM requirements, KV cache overhead, GPU topologies, and power budgets for Llama-3, DeepSeek, and Mistral.
• Dynamic Architecture Blueprint (/blueprint/{token}): Generates personalized enterprise deployment plans on the fly.

4. Cryptographic Air-Gapped Licensing & Buyer Locker

• Issues HMAC-SHA256 signed software license certificates (.lic) for purchased suites.
• Buyer Locker (/buyer) allows verified operators to view entitlement portfolios and download signed update bundles.

---

🚀 Quickstart

1. Local Launch

Windows PowerShell:

.\run.ps1

Linux / macOS:

./run.sh

Server runs on http://127.0.0.1:8020.

2. Verify Fleet Telemetry & Stack Bridge

python scripts/simulate_stack_bridge.py

3. Run Automated Tests

uv run --with pytest --with httpx pytest

Result: 105 / 105 tests passing (100% pass rate).

---

📜 License & Sovereign Stack Heritage

Part of the Hardonia Sovereign AI Operating Stack. Built for independent operators who demand verifiable execution, complete local ownership, and zero data leakage.