Loading…
← All results// Owner ecosystem · live proof
The system behind the architect
8 repos · 1 supervised system · process proof
Before I sell automation to SMB clients, I run it on my own business. This map shows how governance, commerce, AI execution and supervision connect — patterns you can get scaled to your operation, not a ZZP product catalog.
Full architecture map
Adapted from the portfolio ecosystem diagram — Quietforge tokens, B2B framing. Forward to a developer, partner or accountant.
Eight repos at a glance
7 · Method / Automation Map
flexgrafik-meta
Method / Automation Map — the paid first step.
Automation Map sample ↓
How it flows
01
Entry
flexgrafik.nl — portal + chat assistant routes visitors to the right path.
02
Revenue
zzpackage wizard — 7 steps, open pricing, calm order form or designer handoff.
03
Leads
app.flexgrafik.nl game — coupons and contacts funnel into the wizard.
04
Operations
jadzia-core — orders, automation, jadzia.db. Modules refined in agent loop.
05
Execution
Agent OS — Planner → Coder → Tester → Reviewer → Summarizer. Human gate before deploy.
06
Supervision
Flex-VCMS — scan all repos, detect conflicts, enforce SSoT. Governance layer.
// Supervision layer
Why VCMS matters
VCMS is the governance layer behind this ecosystem — not a CMS panel. It scans content, repos and agent rules so your operation does not depend on memory, screenshots or one developer's inbox.
Governance before deploy
VCMS scans all repos, agent rules and handoffs before changes reach production — not after a client reports a bug.
Conflicts visible early
SSoT mismatches and duplicate tasks surface in scan reports. Live target: Conflicts 0 on every scan.
Human approval stays separate
VCMS supervises the codebase; Agent OS runs the build pipeline. Review gates decide what ships — VCMS does not auto-deploy.
- Repo & content scan PROVEN
- Conflict detection + severity PROVEN
- SSoT registry PROVEN
- Governance audit JSONL DEMO
- KODA read-only assistant DEMO
- Human approval gate DEMO
- SSoT health score PLANNED
Maps to case studies
| Case study | Ecosystem layer |
|---|---|
| 01 The self-running back-office | An agent that reads folders, sorts by intent, drafts replies, and stops at a human approval gate before anything sends. |
| 02 Agent OS — supervised agent workforce | FastAPI + LangGraph + PostgreSQL on an EU VPS. Hybrid routing via WAITING_RUNNER: orchestration and HITL online; git/code work on the dev PC. Mission Control UI for tasks, queue, history and costs. |
| 03 Self-service quote & onboarding | A 9-step configurator with open pricing and payment — qualifies, quotes and books without a phone call. |
| 04 Gamified lead system for Dutch ZZP | Game loop + reward ladder → register gate → five acts → leaderboard season → coupon bridge into self-service quoting. |
| 05 Modernisation + AI assistant for an advisory firm | Site modernisation + a lead-qualifying AI assistant (qualification only, no tax advice) + a human-approved content engine, with a signed data-processing agreement. |
| 06 Eight repos, one supervised system | VCMS scans all repos, detects SSoT conflicts, and enforces handoffs before changes reach production. |
Honest framing
No fabricated MRR or uptime percentages. Module status varies by repo — this map shows architecture and process proof, not a promise that every node is flawless. VCMS target: conflicts: 0 on every scan.
Want this level of structure?
You rarely need the full stack on day one. Start with a paid Automation Map — we score which layers matter for your business first.
Book your Automation Map