Priority summary (impact × difficulty ratio)
| # | Gap | Impact | Difficulty | Priority |
|---|---|---|---|---|
| 1 | No single keystone event | High | Medium | 🔴 Phase 1 |
| 2 | Signup required before aha moment | High | Low | 🔴 Phase 1 |
| 3 | Marketplace not a visible growth engine | High | High | 🟠 Phase 2 |
| 4 | MPP Gateway not a narrative wedge | High | Medium | 🔴 Phase 1 (accelerated from Phase 3 per Addendum 2) |
| 5 | No public metrics flywheel | Medium | Very low | 🔴 Phase 1 |
| 6 | No opinionated demo stack (14 frameworks) | Medium | Medium | 🟠 Phase 2 |
| 7 | Activation data not publicly acted on | Medium | Low | 🔴 Phase 1 |
| 8 | Weak creator / community infrastructure | Medium | High | 🟡 Phase 2-3 |
| 9 | No integration with OpenAI / Anthropic / LangChain / top AI coding tools | High | Medium | 🔴🔴 Accelerate to next 30 days |
| 10 | No agent-specific pricing (per-agent spend caps) | Medium | Medium | 🟠 Before MPP mainnet scale |
Gap 1 — No single keystone event (high × medium)
The funnel does not concentrate around one measurable, shareable event. Every breakout competitor organized their entire growth around one such event: - Supabase: database initialization. - Bolt.new: prompt-to-live-app in 30s with no signup. - Lovable: natural language to full-stack working app. - Replit: AI agent builds first app. - Hugging Face: find/use/share a model.
CreateOS has templates, marketplace, deployment, agent payments, MPP, skills, sandbox — a product-completeness strength, a distribution flaw.
Fix: pick one. Recommended: "deploy an earning agent" (see CreateOS - Positioning and Wedge).
Gap 2 — Signup before aha moment (high × low)
Public marketing almost certainly requires signup before the core product is demo-able. Every breakout competitor has moved to zero-signup top-of-funnel. A no-signup demo with a shareable URL and a "claim this project" upgrade path is weeks of work, not quarters.
Gap 3 — Marketplace isn't a visible growth engine (high × high)
Three conditions for marketplaces to become growth engines (Lovable/Hugging Face/Railway pattern): 1. Creators make real, publicly-visible money. 2. Top creators are treated as partners and amplified. 3. Best listings become shareable content on their own.
CreateOS marketplace exists but isn't doing these. Fix: public creator leaderboard with earnings, featured creator program with co-marketing, app-showcase content engine.
Gap 4 — MPP Gateway not a narrative wedge (high × medium) — accelerated
MPP is the most structurally defensible piece of the stack. Currently positioned as a product feature; should be the top-of-stack enterprise narrative. Every enterprise conversation should lead with "CreateOS is where autonomous agents deploy, transact, and earn" — not with compute or deployment. Originally Phase 3; Addendum 2 moved this to Phase 1.
Gap 5 — No public metrics flywheel (medium × very low)
Anton Osika tweeted Lovable numbers daily. Eric Simons published Bolt's "$0 to $4M ARR in 30 days" posts. Amjad Masad ran a public-numbers cadence at Replit. CreateOS has real numbers (credits distributed, deployments, agent MPP transactions, creator payouts) that are currently internal. Weekly public-metrics post is essentially free distribution. Start Tuesday of week 2.
Gap 6 — No opinionated demo stack (medium × medium)
CreateOS supports 14 frameworks. Lovable picked one (React + Tailwind + Supabase), fine-tuned the agent for that stack, and let every other use case fail. This single decision was the most important one Lovable made — and it drove the consistency that made their demos shareable on social. Pick a primary demo stack (recommended: Next.js + Tailwind + PostgreSQL) and make it the default marketing surface. Continue supporting others quietly.
Gap 7 — Activation data not publicly acted on (medium × low)
Activation is documented in memory as the primary constraint. Amplitude charts exist. What's unclear: whether weekly owners are assigned to specific metrics with targets + interventions. RevOps discipline gap, not a data gap. Fix is weekly rhythm with owner + target per metric.
Gap 8 — Creator and community infrastructure (medium × high)
Slow-compounding, most durable channel. Minimum viable program: - Weekly community showcase of user-built apps (Twitter). - Monthly creator interview series. - Quarterly hackathon with marketplace prizes + co-marketing partner (Supabase, Anthropic, etc.).
Gap 9 — No explicit integration wedge with top AI coding tools + SDKs (high × medium) — 🔴🔴 accelerate
This is the single highest-leverage BD move CreateOS can make in the next 30 days.
CreateOS is not on the OpenAI Agents SDK default sandbox provider list (Blaxel, Cloudflare, Daytona, E2B, Modal, Runloop, Vercel). Supabase won partly by being the easiest integration target for Cursor/Bolt/Lovable/v0 — CreateOS's equivalent move is to be the default deployment + monetization target for those tools, via an MCP server + "Deploy to CreateOS" button.
Concrete targets (from CreateOS - BD Targets): - Karan Sharma (OpenAI, Agents SDK product). - Rohan Varma (OpenAI, Codex). - Harrison Chase (LangChain CEO). - Andrew Reed (LangChain DevRel). - Sri Viswanath (Sycamore CEO). - Jansen Teng (Virtuals Protocol).
Gap 10 — Pricing is not agent-priced (medium × medium)
Current: $1 = 100 credits with subscription + top-up. Rational developer pricing, not rational agent pricing. Agents have unbounded usage and need per-agent spend caps + budget isolation. Already flagged in memory as needed before MPP mainnet scale.
Risk callout
Replit's AI agent deleted a client's database in July 2025. This is now a known failure mode of the vibe-coding paradigm. As MPP volume grows, blast radius grows. Ship per-agent spend caps before any major consumer launch. See CreateOS - Risks and Mitigations.
Sources
- CreateOS_Growth_Playbook Part 3
- CreateOS_Playbook_Addendum2_Agent_Infra — accelerated priorities
- CreateOS_Playbook_OnePage_Reference — gap priority order