Roadmap

90 days to launch.

Three phases: build the spine, earn the trust, go public. Every week has a theme, top-3 outcomes, risks, and dependencies.

W1–W4

Foundation

In progress

W5–W8

Closed Beta

Upcoming

W9–W13

Public Launch

Upcoming

90-Day Plan (W1–W13)

Phases: Foundation (W1–W4) → Closed Beta (W5–W8) → Public Launch (W9–W13).

Owners (role tags — assign actual names in your project-management tool once team is confirmed):

`founder` — Jah (iojah@hotmail.com)

`eng-lead` — runtime, ai-hub, plugins

`max-lead` — Max persona/runtime/voice

`web-lead` — `01-website-ai/`

`growth-lead` — `03-marketing-launch/`

`monetization-lead` — `04-monetization/`

`infra-lead` — `08-deployment-ci/`, `05-observability/`

`qa-lead` — `09-tests-qa/`

`legal-lead` — `06-legal-compliance/`

`community-lead` — community channels, partner DMs

Weekly cadence: Monday standup (25 min), Wednesday demo (30 min), Friday metrics (30 min). All anchored on the cells below.

---

Phase 1 — Foundation (W1–W4)

W1 — "Ship the spine"

Theme: Repo, infra, and Max are real and running on a real domain.

Top 3 outcomes:

1. Domain `metatrix.ai` live with `01-website-ai/index.html` deployed to Cloudflare Pages or Vercel via `08-deployment-ci/scripts/deploy.sh`. (`web-lead`) 2. Max runtime `02-max-integration/max-runtime.js` boots on `metatrix.ai/lab` with chakra physics + breath cadence rendering at 60fps. (`max-lead`) 3. Observability baseline: structured logs from `05-observability/server/lib/logger.js` flowing to a hosted sink (Axiom or Grafana Cloud), error budget defined. (`infra-lead`)

Risks: WebXR perf on mid-tier laptops; Max voice latency over public network.

Dependencies: Domain ownership confirmed; secrets vault chosen (1Password or Doppler).

W2 — "Trust contract"

Theme: BYOK + audit chain visible to a real outside user.

Top 3 outcomes:

1. BYOK flow end-to-end: paste key → AES-256-GCM at rest → first call hits user's own provider. Tested with multiple external providers. (`eng-lead`) 2. Enable preview deploys (`08-deployment-ci/.github/workflows/preview-deploy.yml`) so every PR gets a `pr-N.metatrix.ai` URL within 90s. (`infra-lead`) 3. Privacy + ToS v1 published (`06-legal-compliance/PRIVACY_POLICY.md` + new `TERMS.md` + `AUP.md`), reviewed by outside counsel. (`legal-lead`)

Risks: Counsel turnaround > 1 week; provider API breaking changes.

Dependencies: Choice of cookie banner library (Klaro vs custom).

W3 — "Max can do something useful"

Theme: Max ships one full intention → artifact loop.

Top 3 outcomes:

1. Sample project "Build a landing page in 8 minutes" runs end-to-end via `07-onboarding/samples/landing-page.json` with Max narrating. Recorded as a 3-min video for `03-marketing-launch/DEMO_SCRIPT.md`. (`max-lead`) 2. Plugin gallery v1 surfaces top 12 of the 127 plugins with one-click install. (`eng-lead`) 3. Status page (`05-observability/status-page/`) live at `status.metatrix.ai`, scraped by uptime checker. (`infra-lead`)

Risks: Demo flake on stage; plugin metadata inconsistency.

Dependencies: Decision: which 12 plugins are flagship (see `OPEN_QUESTIONS.md` Q5).

W4 — "Pricing is real"

Theme: Payments live, four tiers visible, one paid customer (founder OK).

Top 3 outcomes:

1. Payments wired via `04-monetization/server/routes/checkout.js`; Free / Builder $19 / Studio $79 / Enterprise sales-only live on `01-website-ai/pricing.html`. (`monetization-lead`) 2. Usage meter shipping events from `04-monetization/server/middleware/usage.js` into Postgres + billing meter. (`monetization-lead`) 3. First 25 closed-beta invites sent via `03-marketing-launch/EMAIL_SEQUENCE.md` sequence #1 to seed list. (`growth-lead`)

Risks: Tax registration delays for non-US users; meter overcounts on retried calls.

Dependencies: Decision on annual discount % (see `PRICING_STRATEGY.md`).

---

Phase 2 — Closed Beta (W5–W8)

W5 — "Real strangers"

Theme: 25 outside users in the room, instrumented.

Top 3 outcomes:

1. 25 closed-beta accounts active; activation funnel dashboard live in Grafana from `05-observability/dashboards/activation.json`. Target: 60% reach "first artifact shipped." (`growth-lead`, `infra-lead`) 2. 5 Playwright E2E flows in `09-tests-qa/e2e/` green on every preview deploy: signup, BYOK, build-site, plugin-install, audit-export. (`qa-lead`) 3. First weekly NPS survey shipped (`07-onboarding/samples/feedback-survey.json`), n ≥ 15 responses. (`growth-lead`)

Risks: Activation < 30% (kills the launch story); high BYOK abandonment at the paste step.

Dependencies: a community channel chosen for the beta cohort.

W6 — "Max stops being weird"

Theme: Persona stability + the "creepy/cringe" risk neutralized.

Top 3 outcomes:

1. Persona regression suite (`02-max-integration/tests/persona.spec.js`) with 50 prompts × 5 personas; ≥ 95% on-voice rate. (`max-lead`, `qa-lead`) 2. A/B persona variants behind a feature flag (LaunchDarkly or Unleash); see `AGENT_OPS_PLAN.md`. (`max-lead`) 3. 3 unscripted user reactions filmed on Mac webcam, edited to 30s clips. If anyone says "creepy" we re-tune. (`growth-lead`)

Risks: R-22 in `RISK_REGISTER.md` (Max perceived as creepy / cringe).

Dependencies: Filmed-consent waiver in `06-legal-compliance/`.

W7 — "Plugin gravity"

Theme: Make the plugin/MCP story undeniable.

Top 3 outcomes:

1. MCP server interop with third-party MCP clients verified end-to-end; recorded demo. (`eng-lead`) 2. Five named outside plugin authors shipped a plugin to the gallery (see `PARTNERSHIPS.md` creator tier). (`community-lead`) 3. Plugin SDK docs live at `metatrix.ai/docs/plugins`, source from `01-website-ai/` + dev-portal extension. (`web-lead`, `eng-lead`)

Risks: Plugin auth surface (R-9 BYOK breach via malicious plugin).

Dependencies: Plugin sandboxing model decision (see `OPEN_QUESTIONS.md` Q11).

W8 — "Beta → green"

Theme: Hit the launch-readiness gate or push public launch one week.

Top 3 outcomes:

1. All P0 items from `09-tests-qa/` perf + security + chaos suites passing; uptime in last 14 days ≥ 99.5%. (`qa-lead`, `infra-lead`) 2. 50 active beta users; activation ≥ 50%, week-2 retention ≥ 35%. (`growth-lead`) 3. Launch checklist (`03-marketing-launch/LAUNCH_CHECKLIST.md`) at 90% green; founder reviews red items personally. (`founder`)

Risks: Hidden P0 surfaced at the last minute; press timing slipping.

Dependencies: Launch-day infra capacity plan (see W9).

---

Phase 3 — Public Launch (W9–W13)

W9 — "Pre-launch press"

Theme: Story is in the hands of journalists and creator partners.

Top 3 outcomes:

1. Embargoed press kit sent to 12 outlets (TechCrunch, The Verge, Ars Technica, Wired, The Information, Latent Space, The Pragmatic Engineer, Ben's Bites, TLDR AI, Lenny's Newsletter, Hacker News submission accounts queued, Product Hunt scheduled). See `03-marketing-launch/PR_OUTREACH.md`. (`growth-lead`) 2. Three creator partners (named in `PARTNERSHIPS.md` creator tier) have working accounts and pre-recorded reactions. (`community-lead`) 3. Capacity: load test from `09-tests-qa/perf/` shows system handles 10k concurrent + 1k/min signups; Cloudflare or equivalent rate limits configured. (`infra-lead`)

Risks: Embargo break; one creator backs out 48h before.

Dependencies: Founder voice on Latent Space or Lex Fridman scheduled or pitched.

W10 — "Launch week"

Theme: Product Hunt + Hacker News + email + social, in that order.

Top 3 outcomes:

1. Product Hunt #1 of the day or top 3. (`growth-lead`, `community-lead`) 2. Hacker News post titled per `03-marketing-launch/LAUNCH_NARRATIVE.md` lands in top 5 for at least 4 hours. (`founder`) 3. 1,000 new signups inside the launch week; activation ≥ 45%. (`growth-lead`)

Risks: HN flame-out on Max's framing (see R-22); Payment webhook backpressure under signup spike.

Dependencies: Pre-warmed CDN; on-call rotation defined (see `05-observability/alerts/runbooks/`).

W11 — "Convert beta to paid"

Theme: Free → paid funnel proves out.

Top 3 outcomes:

1. Paywall metering from `04-monetization/server/lib/limits.js` enforces Free tier; nudge surfaces at 80% usage. (`monetization-lead`) 2. 5% free → paid conversion in launch cohort by end of week. (`monetization-lead`, `growth-lead`) 3. First five Studio-tier ($79) self-serve customers without a sales call; case studies queued. (`monetization-lead`)

Risks: Paywall annoys early advocates; refund spike.

Dependencies: `BYOK_BILLING_POLICY.md` published; `INVOICE_TEMPLATE.md` rendered correctly for EU VAT.

W12 — "Enterprise lane opens"

Theme: First enterprise pilot signed.

Top 3 outcomes:

1. Enterprise pricing page (`01-website-ai/enterprise.html`) live with self-host CTA; DPA + SOC 2 roadmap published. (`web-lead`, `legal-lead`) 2. One signed enterprise pilot LOI (target: dev-tools company or creator collective with > 50 seats). (`founder`) 3. SSO (SAML or OIDC) wired via Clerk or WorkOS for the pilot. (`eng-lead`)

Risks: Enterprise procurement length kills momentum; no SOC 2 yet.

Dependencies: SOC 2 Type 1 prep started (see `OPEN_QUESTIONS.md` Q19).

W13 — "Reflect, re-rank, re-plan"

Theme: Score the launch and write Q2.

Top 3 outcomes:

1. OKR scorecard for Q1 published in `OKRS.md`; Q2 OKRs drafted. (`founder`) 2. Launch retro: top 5 wins, top 5 misses, 10 lessons; persisted in `OPEN_QUESTIONS.md`. (all) 3. Re-rank `ICE_BACKLOG.md` based on real activation/retention data; pick next sprint top-10. (`eng-lead`, `founder`)

Risks: Founder fatigue; over-correcting on one viral comment.

Dependencies: Real numbers from `KPI_DASHBOARD.md` over 14 days post-launch.

---

Cross-week disciplines (don't skip)

Daily: error-budget burn check (`05-observability/dashboards/sli.json`).

Weekly: persona regression run (`02-max-integration/tests/`); chaos drill once a week from `09-tests-qa/chaos/`.

Bi-weekly: prune `ICE_BACKLOG.md`; re-score top 20.

Monthly: review `RISK_REGISTER.md`, update triggers fired, add 2 new risks.

Failure modes that should re-trigger this plan

Activation < 30% at end of W5 → pause launch, reset W5–W8, run an onboarding sprint.

Beta NPS < 20 at end of W7 → escalate persona to founder, consider pulling Max from default and putting it behind a feature flag (loss of differentiation, but salvageable).

Payments live-mode incident → freeze monetization changes, reverse to W4 baseline.

Press leak before W9 → fold W9 forward into the next available week, do not chase coverage.