Current Portfolio
What Nova AI Ventures is building right now. 15 active product positions across L02a/L02b · 1 L03 SMB template · 3 L04 AI Management Suite deployments · 1 L05 Venture R&D engagement.
1. Portfolio Overview Dashboard
📦 CURRENT PORTFOLIO · OVERVIEW
Stages: POC (working proof of concept) · MVP (minimum viable product · staging live) · Published (live in production · users can install/use). Pre-POC means design phase · not yet built.
📦 PORTFOLIO · 14 PRODUCTS
SnapSell
L02a · Published · C2C marketplace
My Story Machine
L02a · POC · children’s publishing
Iris AI
L02a · POC · art domain
Bartender
L02a · POC · consumer hospitality
Comic Chef
L02a · POC · consumer creator
Family Budget
L02a · POC · consumer finance
Paddle Analyzer
L02a · POC · consumer sport
Vet Snap
L02a · POC · pet / B2B vet
PropertyFinder
L02a · POC · real estate
Knowledge Link
L02a · POC · consumer knowledge
Legal Runner
L02a · POC · legal tech
PA2Digital
L02a · POC · consumer AI admin
Exam Prep Tool
L02a · POC · knowledge tester
Lumen
L02b · POC/MVP · wealth mgmt (Ross Smith)
Otherth.
L02b · POC · B2B marketing platform
CNC RadFres
L03 · Pre-POC · SMB manufacturing template
NAIV Internal
L04 · Live · internal dogfooding
Codibly Case Study
L04 · Pipeline · first external deployment
RadFres Case Study
L04 · Pipeline · combined SMB + AI Mgmt Suite
Softiq
L05 · Active · enterprise R&D (Extract + Validate)
1.1 · L02a · Apps · Nova AI-initiated (13 products)
| Product | Category | Stage |
|---|---|---|
| SnapSell | C2C marketplace | Published |
| My Story Machine (aka Albik Creator) | Children’s publishing · family story printable books · talking pen apps for families and children | POC |
| Iris AI | Art domain | POC |
| Bartender | Consumer hospitality | POC |
| Comic Chef | Consumer creator | POC |
| Family Budget & Investment Tracker | Consumer finance | POC |
| Paddle Game Real-time Analyzer | Consumer sport | POC |
| Vet Snap | Consumer pet / B2B vet | POC |
| PropertyFinder | Consumer real estate | POC |
| Knowledge Link App | Consumer knowledge | POC |
| Legal Runner | Consumer legal tech | POC |
| PA2Digital Assistant | Consumer AI admin | POC |
| Exam Preparation Tool (Knowledge Tester) | Consumer knowledge tester | POC |
1.2 · L02b · Apps · Co-founded with expert (2 products)
| Product | Partner | Stage |
|---|---|---|
| Lumen Wealth Management | Ross Smith (idea originator + product owner + investor) | POC/MVP |
| Otherth. Marketing Platform | Partner → future CEO | POC |
1.3 · L03 · Productized SMB Implementation (1 template in design)
| Template | Pattern | Stage |
|---|---|---|
| CNC RadFres | Manufacturing / CNC ops | Pre-POC (design phase) |
1.4 · L04 · AI Management Suite (productized)
| Deployment | Status |
|---|---|
| Nova AI Ventures itself | Live · internal dogfooding |
| Codibly case study | Pipeline · potential first external deployment |
| RadFres case study | Pipeline · combined SMB + AI Management Suite deployment (moved from L05) |
1.5 · L05 · Venture R&D (1 active)
| Engagement | Filter test | Status |
|---|---|---|
| Softiq | Extract + Validate | Active enterprise implementation engagement |
Line transition rule: L05 Venture R&D engagements that successfully produce reusable platform assets may transition into L03 templates or L04 deployments post-delivery. Classification reflects current phase, not fixed category.
2. Nova AI Stack Reuse Matrix
🔄 STACK REUSE · WHAT REUSES WHAT
Estimated reuse per product per layer (L01-L06). High-level %.
| Product | L01 Platform | L02 Backend | L03 LLM Gateway | L04 AI Orch | L05 Admin UI | L06 Surfaces | Overall |
|---|---|---|---|---|---|---|---|
| SnapSell | source | source | source | source | source | source | source |
| Albik Creator | 95% | 90% | 85% | 75% | 80% | 60% | ~85% |
| Iris AI | 95% | 90% | 85% | 75% | 80% | 55% | ~80% |
| Bartender | 95% | 90% | 85% | 70% | 80% | 50% | ~75% |
| Comic Chef | 95% | 90% | 85% | 70% | 80% | 50% | ~75% |
| Family Budget | 95% | 90% | 85% | 75% | 80% | 60% | ~80% |
| Paddle Analyzer | 95% | 85% | 80% | 60% | 75% | 45% | ~70% |
| Vet Snap | 95% | 90% | 85% | 75% | 80% | 60% | ~80% |
| PropertyFinder | 95% | 90% | 85% | 80% | 80% | 65% | ~85% |
| Knowledge Link | 95% | 90% | 85% | 75% | 80% | 60% | ~80% |
| Legal Runner | 95% | 90% | 85% | 75% | 80% | 60% | ~80% |
| PA2Digital | 95% | 90% | 85% | 80% | 85% | 65% | ~85% |
| Lumen | 95% | 95% | 90% | 75% | 80% | 50% | ~80% |
| Otherth | 95% | 90% | 85% | 75% | 85% | 70% | ~85% |
| CNC RadFres (L03) | 95% | 90% | 85% | 70% | 90% | 40% | ~80% |
| RadFres SMB+AI | 95% | 90% | 85% | 80% | 90% | 50% | ~85% |
| AI Mgmt Suite internal | 95% | 90% | 90% | 95% | 95% | 75% | ~90% |
Pattern observations
- L01 Platform reuse stable at 95% — Google/Cloudflare baseline applies universally
- L02 Backend reuse stable at 90% — Docker/Traefik/Sentry universally reusable
- L03 LLM Gateway stable at 85-90% — LiteLLM Proxy pattern universal; only token branding + guardrails product-specific
- L04 AI Orchestration variable (60-95%) — depends on domain-specific agents + MCPs
- L05 Admin UI stable at 75-90% — template scaffolding universal; dashboards product-specific
- L06 Surfaces highly variable (40-75%) — product UI is the biggest differentiator
Average Nova AI Ventures portfolio reuse: ~80-85% from Stack · 15-20% product-specific work
3. SnapSell Tax Payoff Table
🧾 SNAPSELL TAX PAYOFF · 5 MONTHS OF PAIN
SnapSell zapłacił 5 months (2.25× over plan) · każdy następny produkt dostaje inheritance free.
SnapSell · The Full History
SnapSell is Nova AI Ventures’ first live product — C2C marketplace AI for photo-to-listing on Vinted and OLX. Shipped by Bart as a test-drive on Google Cloud Run on November 25, 2025. Maciek said “LOVE IT.” We said launch in two months. We ended up five months in before going live.
Timeline
- Nov 25, 2025 — First test-drive on Google Cloud Run · marketing video shipped · “Launch in 2 months”
- Jan 2026 — Vinted API ruled out · OLX OAuth investigated · Maciek’s competitive research lands
- Feb 24, 2026 — Petro Polishchuk joins as contract dev (~15k PLN/mo)
- Mar 4–13, 2026 — OLX OAuth flow + Vinted integration shipped · admin panel + token economy · latency 81s → 12s
- Mar 22, 2026 — Vincent posts first real OLX listing using the app — Radlands board game, 220 zł · “I am impressed. Well done guys.”
- Mar 24–31, 2026 — First native iOS + Android builds · IAP + Google Play Billing · Sentry · Apple Developer + Play Console bootstrap
- Apr 25, 2026 — Both iOS and Android published in stores · pending friends & family launch + Vincent marketing activation
What we learned (beyond the 3 lessons below)
- API due diligence before committing. Vinted rejected. OLX approved. Allegro investigated. This was done in Jan 2026 before committing to build — saved us from a dead-end integration.
- Legal work belongs at Day 1, not Day 60. T&C, Privacy, Refund, Cookie templates started Jan 21, 2026. By the time the store submission happened in April, legal was ready. If we had started legal later, we’d have missed the window.
- Vincent shifted from code to marketing Feb 17. “Don’t want to do more code, focus on marketing/distribution.” This is the moment the CGO role fully crystallized. Future products should expect founder roles to crystallize mid-build — do not hold rigid role assignments from Day 0.
- Payment integration is non-trivial. Stripe + Apple IAP + Google Play Billing. Three separate integration paths. Each has its own review process. Budget 2–3 weeks for payment plumbing per product.
- Analytics came in late — should come earlier. Amplitude + Meta Pixel + Google Ads were added April 17, after store submission. Next product: analytics designed into the build from G1, not retrofitted.
- Operator search never happened. Five months in · still no named lead · Vincent is the internal operator by default. This is the biggest gap — validates §09 rule that operator search must start at G1.
- Dual-platform launch pattern works. Simultaneous iOS + Android submission was the right call. Apple approved fast (48 hours) · Google took ~7 days. The 7-day Google wait is now a known planning constant.
- Wave-based delivery discipline emerged gradually. Early SnapSell work was not wave-based. Dexter delivery doctrine (§07) crystallized as a response to SnapSell being too loose on QA. Every next product starts wave-based from Day 1.
What SnapSell paid for · what every next product inherits
| Category | SnapSell paid | Universal inheritance | Conditional inheritance |
|---|---|---|---|
| Legal | T&C, Privacy, Refund, Cookie templates (Polish law) | ✅ All PL-market products | Non-PL products need regional versions |
| Payments | Stripe web + Apple IAP + Google Play Billing patterns | ✅ All monetized products | Free products skip IAP |
| Observability | Sentry cross-platform + OpenTelemetry | ✅ All products | — |
| LLM Gateway | LiteLLM Proxy per-product instance pattern | ✅ All products | Token naming varies (Cook Coins, Lumen Credits, etc.) |
| OAuth integrations | Google OAuth · Apple OAuth · OLX OAuth patterns | ✅ Google + Apple OAuth (every product) | OLX-style third-party OAuth (marketplace products like PropertyFinder, Vet Snap) |
| Admin panel | Token economy dashboard + per-item cost/margin visibility | ✅ All products | Product-specific metrics layered on top |
| Hosting | Google Cloud Run + Firebase hosting patterns | ✅ All products | — |
| Mobile builds | Native iOS + Android pipeline · Apple Developer · Google Play Console | ✅ All mobile products | Web-only products skip |
| Product Builder | 7-stage pipeline + folder scaffolding + gate system | ✅ All products | — |
| Dexter delivery doctrine | Wave-based · Playwright · Vera loop | ✅ All products post-G4 | — |
| CI/CD | App Store Deployment Claude Skill · build automation | ✅ All mobile products | — |
SnapSell Lessons Applied (from §02)
Every product in portfolio starts with:
- ✅ Lesson 01: Stand framework up first · clean wireframes before prototyping
- ✅ Lesson 02: Foundation exists before build starts (no “build runway while taking off”)
- ✅ Lesson 03: Starts at month 2, not month 0
4. Active Pipeline · Condensed Summary
🚀 ACTIVE PIPELINE · WHAT'S NEXT
A compact view of every active position. One sentence per product — what it is and what state it’s in. For deep-dives on flagship products (SnapSell · Lumen · Otherth · My Story Machine) the full conversation lives in the project Slack/Telegram channels per §07 Stage 2.
L02a · Apps · Nova AI-initiated
- SnapSell — C2C marketplace AI (photo-to-listing for Vinted/OLX); published in both stores Apr 2026; pending friends & family launch and Vincent marketing activation.
- My Story Machine (aka Albik Creator) — children’s publishing + family story printable books + talking-pen apps for families and children; POC stage, operator search active.
- Iris AI — art-domain creative tooling; POC stage, domain scope still being sharpened.
- Bartender — consumer hospitality tooling (bar/cocktail workflow); POC stage.
- Comic Chef — consumer creator app for comic-style content; POC stage.
- Family Budget & Investment Tracker — consumer finance for high-income busy families (distinct from Lumen’s wealth-management segment); POC stage.
- Paddle Game Real-time Analyzer — consumer sport app for paddle match analysis; POC stage.
- Vet Snap — consumer pet / B2B vet marketplace-style app; POC stage.
- PropertyFinder — consumer real estate marketplace (heavy reuse from SnapSell OLX patterns); POC stage.
- Knowledge Link App — consumer knowledge / productivity RAG-heavy app; POC stage.
- Legal Runner — consumer legal tech (Mikołaj’s domain fit); POC stage.
- PA2Digital Assistant — consumer AI admin assistant (productized version of AI Management-Suite patterns); POC stage.
- Exam Preparation Tool (Knowledge Tester) — consumer knowledge-testing / study aid app; POC stage.
L02b · Apps · Co-founded with expert
- Lumen Wealth Management — AI-powered Family Financial OS for $2M–$20M net worth segment; co-founded with Ross Smith (idea originator + product owner + investor); POC/MVP stage.
- Otherth. Marketing Platform — B2B marketing platform; co-founded with partner who is lined up as future CEO; POC stage, partnership terms in discussion.
L03 · Productized SMB Implementation
- CNC RadFres — manufacturing / CNC ops template (first SMB implementation candidate); pre-POC design phase.
L04 · AI Management Suite (productized)
- Nova AI Ventures itself — live internal dogfooding of the 11-agent Management Suite that we also sell externally.
- Codibly case study — pipeline: first external L04 deployment · strategic opportunity given Bart’s relationship · COI disclosure handled when going public.
- RadFres case study — pipeline: combined SMB + AI Management Suite deployment (moved from L05 · this is deployment of a productized capability, not pure R&D).
L05 · Venture R&D
- Softiq — active enterprise implementation engagement; filter passed on Extract + Validate tests; reusable enterprise MCP patterns as the asset deliverable.
Line transition rule: L05 Venture R&D engagements that successfully produce reusable platform assets may transition into L03 templates (e.g., CNC RadFres pattern) or L04 deployments (e.g., RadFres case study, which did exactly this). Classification reflects current phase, not fixed category.