We help creators know why their video works — or how to fix it — before they publish. The craft stays theirs.
Eight steps. An hour per Reel-post. I ran the manual chain — download MP4, transcribe, paste into Claude, frame, refine, post — for months before I built a script to skip it. Within weeks, strangers were finding my tool through ChatGPT. That's when I knew this wasn't my pain alone — it was a category.
Solo founder · ex-CTO Neural Dynamics
Co-founded and scaled Neural Dynamics from 3 to 25 engineers as CTO. Shipped 30+ platforms across SaaS, CRM, ERP, AI, fintech for US, EU, and Asia clients. Built Lomero solo in 8 weeks — full product, open signup, organic AI-search distribution.
Two recent AI workflows I've shipped to enterprise clients — selected to show range across real-time pipeline ops and document automation:
Veo 3, Sora 2, Runway, Luma. Text-to-video down ~80% in 18 months. Within 12 months, generating a short-form video will cost less than a text message.
Netflix CCO runs $17B budget. PayPal hired Head of Content inside CEO's office. 419 open Director/Chief Content Officer roles at $120-230K base — fintech and non-media included.
Viraliq iOS-only, dying. TikAlyzer 5 organic visits/mo. VidIQ / TubeBuddy YouTube-locked. No incumbent owns cross-platform + web + pre-publish + AI-aware.
The sequence flipped. Creators build audience first, then build the business around it. Content stopped being a marketing function — it became the acquisition engine.
The industry calls it the same way






How we deliver it · four beats
Product stack
| Layer | Status | What it ships |
|---|---|---|
| L0 | ✓ Live | Pre-publish breakdown · URL paste · Reel/TikTok/Short · hook + structure + pacing + CTA scored · niche-matched references |
| L1 | Phase 1 · Q1 | Prescription Engine · 3-5 hook variants + shot list + b-roll checklist · niche × tier matched · plugs into Higgsfield / Runway / Pika |
| L3 | Phase 2 · Q2 | Closed-loop validator · brief vs shot video deviation · outcome capture · per-user, per-niche compounding dataset · THE MOAT |
Cost curve · per-analysis COGS
| Stage | When | Per-analysis COGS | Gross margin · Creator $19 / 50 analyses |
|---|---|---|---|
| Stage 1 · Third-party APIs (Claude + Whisper + Perplexity) | Today | ~$0.10 | ~74% |
| Stage 2 · RAG-optimized + caching | Ships M3-M4 | ~$0.03 (-70%) | ~92% |
| Stage 3 · Local GPU + fine-tuned model | Year 2 | ~$0.008 (-70%) | ~98% |
Data flywheel
Pattern library across niches, formats, creator tiers. Re-creatable by competitors with 6 months + budget. Table stakes. Not the moat.
Our brief → user shoots → real-world result. Per-user, per-niche. Compounds with every active user. Nobody else has this. Can't be scraped retroactively. Platforms won't build it — conflicts with their ad model.
Direct competitors (claim our slot)
| Competitor | Position | Stage | Threat |
|---|---|---|---|
| Sandcastles | "Find outliers + remix + script generator." Database + script + hook generation. MCP plugin live, public API, $39-399/mo. | ~105K signups claimed (real paid ~1-5K). v2 shipped. | Highest |
| TransClipper | "Uncover hook, structure, CTA behind any video." Multi-video pattern detection. | Cited #1 by ChatGPT on commercial queries. | High — third-party authority signal |
| ViralWatch | "Score-backed analysis with concrete edits." | Early. | High — closest tagline copy to our wedge |
iOS-only / failed: Viraliq (0 organic, 1 Meta ad, dying) · TikAlyzer (5 organic, validation-mode ads, no traction). Adjacent extension risks (watching): VidIQ / TubeBuddy (YT-locked) · Opus Clip (generator) · Submagic / CapCut / Descript (editors) · Higgsfield (12-sec hook scorer) · Memories.ai (enterprise, wrong ICP).
Why a 2026 product: Generation cost only crossed the "free" threshold in 2025 · AI search (ChatGPT, Perplexity) now refers users to tools they don't know exist · Content leadership became a C-suite function in the last 18 months.
Each tier unlocks the next. Free tier serves the belief — everyone deserves a real shot at building a brand. Paid tiers come from creators with commercial intent.
~27 organic from marketing · 3-5 new signups / day
Traffic mix · last period · 567 sessions
AI search refers users who didn't know we existed — the moat starting.
Engagement quality
| Category | Q1 | Q2 | Q3 | Q4 | Year |
|---|---|---|---|---|---|
| Founder ($5k/mo) | 15.0 | 15.0 | 15.0 | 15.0 | 60.0 |
| Engineer FT (M2 start) | 10.0 | 15.0 | 15.0 | 15.0 | 55.0 |
| CMO FT (M2 start) | 12.0 | 18.0 | 18.0 | 18.0 | 66.0 |
| Outsource (brand · SEO · UX · data ops) | 10.0 | 4.5 | 4.5 | 4.5 | 23.5 |
| Personnel subtotal | 47.0 | 52.5 | 52.5 | 52.5 | 204.5 |
| Marketing ads | 3.0 | 7.0 | 10.0 | 10.0 | 30.0 |
| Compute + AI APIs | 2.0 | 3.0 | 4.0 | 6.0 | 15.0 |
| Office (Lisbon-band · M3 start) | 1.6 | 4.8 | 4.8 | 3.3 | 14.5 |
| Travel + Partnership | 1.0 | 2.0 | 4.0 | 3.0 | 10.0 |
| Subscriptions (eng + mkt + biz) | 1.5 | 2.0 | 2.5 | 2.5 | 8.5 |
| Equipment (one-time) | 8.0 | — | — | — | 8.0 |
| Legal (incorp + contracts) | 3.0 | 1.0 | 0.5 | 0.5 | 5.0 |
| Accounting | 0.5 | 0.7 | 0.7 | 0.6 | 2.5 |
| Hiring (screening tools) | 2.0 | — | — | — | 2.0 |
| Insurance D&O (M6 start) | — | 0.25 | 0.75 | 0.5 | 1.5 |
| Non-personnel subtotal | 22.6 | 20.75 | 27.25 | 26.4 | 97.0 |
| Buffer / contingency (14%) | 10.4 | 12.75 | 12.25 | 13.1 | 48.5 |
| Quarterly total ($k) | 80.0 | 86.0 | 92.0 | 92.0 | 350.0 |
All figures in $k USD · Cumulative draw: 23% → 47% → 74% → 100% · Avg monthly burn $29.2k · EU base, geo locked at first investor commit
12-month revenue scenarios
| Scenario | Free users | Creator $19 | Pro $49 | Agency $149 | MRR | ARR |
|---|---|---|---|---|---|---|
| Conservative | 5,000 | 100 (2%) | 30 | 5 | $4.4k | $53K |
| Base | 12,000 | 360 (3%) | 144 | 20 | $18.0k | $215K |
| Aggressive | 30,000 | 1,500 (5%) | 750 | 60 | $81.7k | $980K |
| Phase | Timing | Goal | Activation gate | Validation metric |
|---|---|---|---|---|
| 1 MVP (L1) | M1-M3 | Brief Generator from connected accounts | Signup → first brief ≤3 clicks, <90s | >40% generate ≥3 briefs week 1 |
| 2 Closed Loop (L3 basic) | M4-M6 | Upload shot video, compare vs brief | Outcome captured after publish | >25% close loop on 1+ video M1 |
| 3 Series Mode | M6-M10 | 3-5 briefs at once, mutually aware | Justifies Pro $49 tier | >60% Pro users create a series M1 |
| 4 Publishing Integrations | M10-M12 | Buffer / Later / Metricool — demand-gated | Reduces end-of-loop friction | >30% adoption post-enable |
| 5 Tinder Calibration | Conditional | 10-card onboarding for low-history users | Only if Phase 1 inference fails | TBD |
| Risk | What we watch | Mitigation |
|---|---|---|
| Sandcastles ships closed-loop validation | Updates mentioning "track your video performance," "outcome," "before/after," "performance vs prediction." | Ship Lomero MCP plugin in 30-60 days · close-loop dataset ramp gets priority engineer time |
| VidIQ / Opus Clip extends into short-form pre-publish | Adjacent incumbents with distribution. Could ship "score your draft" in 6-12 months. | Cross-platform from day one · depth of scoring · data moat compounding faster than side-feature |
| AI search referral channel dries up | ChatGPT/Claude/Gemini referrals = 28% of traffic. Behavior changes or pay-for-placement could evaporate this. | Diversify: paid acquisition Q2+ · creator partnerships · content SEO · don't be fully AI-dependent |
| Free-to-paid conversion < 1.5% | Plan assumes 2-3% conversion. Below 1.5% = unit economics tight. | Tier tuning: gate features earlier · narrow free tier · raise paid price · Series Mode is the leverage |
| L1 Brief Generator activation < 40% | Phase 1 kill criterion. Users don't generate 3+ briefs week 1. | Re-plan Phase 2 · add Tinder calibration (Phase 6) · harder reference matching |
| CMO hire fails | Senior creator-economy CMOs are rare. Slipped close in M1-M2 = GTM slips. | Direct sourcing only · screening tools earmarked · backup: contract Head of Growth + PT content lead |
| Geo / visa surprises | EU relocation + in-person team. Visa for non-EU founder + hires. | Defer geo final lock until first investor commit · plan in geo-agnostic numbers (Lisbon baseline) |