FocusFlow is an academic operating system for neurodivergent students — an AI study coach, an executive-function toolkit, and a safe tutor marketplace, in one app. It's not a prototype. It's live in production today.
Notion, Quizlet and generic tutoring all assume the hard part is content. For a neurodivergent student, the hard part is starting, staying on track, and self-regulating under pressure — the executive-function layer no other tool owns.
Every module writes to the same student model, so the coach knows what's due, what's stressful, and what a student is weak on — and gets proactive before a test. This is the full surface, all shipped.
A Socratic tutor that adapts to the student's mode (ADHD, autism, exec-function, exam-prep) and refuses to hand over answers — it teaches, it doesn't cheat. Safe for parents and schools.
3-provider engineCommand Centre "mission control," AI task-breakdown into tiny steps, one-tap quick-capture brain dump, Pomodoro focus mode, a Calm/overwhelm flow, and a pre-exam "lockdown" mode.
Command CentreFlashcard decks with spaced repetition, auto-generated quizzes & practice exams, printable guides, an Explain-it-back (Feynman) mode, and an OCR homework scanner.
SRS + FeynmanSchool → term → course → weighted categories, live GPA on an editable scale, and a "what do I need on the final?" calculator that solves the target for you.
Live GPADyslexia-friendly typeface, text scaling, read-aloud TTS, voice dictation, reduced-motion, high-contrast and simplified layouts — applied app-wide from an onboarding neuro-profile.
First-class, not bolt-onTwo-sided marketplace with ID-verified tutors, transparent ranking, a full booking lifecycle, reviews, referrals, and moderation — parent-mediated & safety-gated for under-18s.
Payments built-inBecause school isn't the whole student: mood check-ins, a wellness log, to-dos, and a private budget (income & expenses). Feeds the coach — heavy stress dials back the workload.
Whole-studentA Connect SDK any external study app can drop in. It streams a student's weak topics back in, so the coach nudges practice on exactly what they're struggling with — before the test.
Platform playInstallable PWA with web push and email. Adaptive nudges fire at each student's personal best focus hour, and swap to urgent copy inside a pre-exam lockdown window.
Meets them where they areRetrofitting real neuro-adaptive coaching onto a general study app is a rebuild, not a setting. FocusFlow's data model, coaching modes and UI were designed for this from line one — the incumbents can't fast-follow it cheaply.
It's Socratic by design and redirects answer-seeking. That's the exact property parents and schools require before they'll trust — and pay for — an AI in their child's education. It's a trust asset competitors giving away answers can't claim.
The default AI engine runs locally and free — no API bill, works offline. Paid providers are optional upgrades. Every premium capability (payments, video, alerting) is a dark "seam" flipped on by a single environment variable. Gross margin is structurally high from day one.
Grades, wellbeing, tutoring and daily planning in one place means switching cost compounds. A student's history, streaks, and weak-area signals live here — leaving means starting over.
This isn't "we'll monetize later." A two-sided marketplace with real payments and a $5/mo tutor subscription is coded, tested, and already proven end-to-end against live Stripe — waiting on a go-live decision, not a build.
This reads like a funded team's Series-Seed codebase, built solo in about ten weeks. It's deployed, tested, and monitored. Several revenue and premium features are fully built and sitting one credential away from live.
| Capability | Status | What it takes to switch on |
|---|---|---|
| Core app (coach, study, grades, life) | Live | Running in production on the free local AI engine. |
| Accessibility engine | Live | Shipped and applied app-wide. |
| Email + web-push notifications | Live | Sending in production. |
| Premium AI (Gemini / Claude) | Live | Wired; premium responses enabled via API key. |
| Tutor marketplace + payments | Test mode | Proven end-to-end against live Stripe in test mode. Go-live = swap to live keys + onboard first tutors. |
| Live video tutoring (LiveKit) | Built · dark | Fully coded; needs LiveKit credentials (3 env vars). |
| Error alerting webhook | Built · dark | Sentry live; team alerts need one Slack/Discord URL. |
Not a weekend hack — this is disciplined, testable, deployable software.
Current, mainstream, hireable-for technology — nothing exotic to maintain.
The whole student experience is free — that's the growth engine. Revenue rides on the marketplace and premium tiers, and the plumbing for both is already in the codebase.
Verified tutors pay to appear in discovery. Coded, tested, and validated against live Stripe. Supply-side recurring revenue that scales with the tutor base.
A feature-gating billing seam already exists — every feature currently sits in the free tier. Flip advanced coaching, unlimited premium-AI, or family accounts into a paid plan with no rebuild.
Multi-role support (admin, teacher, class dashboards, moderation, audit logs) is already in place — the foundation for selling seats to schools, districts, and learning-support programs.
Just $5/mo × paying tutors — before any premium student tier, marketplace take-rate, or B2B seats. It shows the shape of the recurring line once supply is on the platform.
| Paying tutors | Monthly recurring | Annual recurring |
|---|---|---|
| 100 | $500 | $6,000 |
| 500 | $2,500 | $30,000 |
| 1,000 | $5,000 | $60,000 |
| 5,000 | $25,000 | $300,000 |
Because the default AI runs locally and free, and every paid dependency is optional, the cost to serve a free user is essentially hosting. Revenue features add cost only when they're already earning. That's a rare position: we can grow the free base aggressively without growing the bill.
The product risk is largely retired. The remaining work is go-to-market and operational, and it's normal for this stage.
The marketplace is pre-launch. We have a proven product but zero usage/retention data. First priority is a real pilot cohort to generate evidence.
Flipping to real charges is a config + compliance step (live Stripe keys, business verification), not a build.
One person built all of this. To scale reliably it needs a small team — the clean, tested, documented codebase makes onboarding realistic.
Currently on Vercel's Hobby plan (crons worked around via GitHub Actions). A modest infra spend unlocks headroom for real traffic.
The product is done enough to test with real students this term. My recommendation is to prove demand before spending on scale.
Recruit 15–30 students (a class, a tutoring group, a support program). Turn on premium AI, watch usage and retention, and gather testimonials. Near-zero cost; produces the evidence every later decision needs.
Swap Stripe to live mode, onboard a first handful of verified tutors, and switch on LiveKit video. Turns a built feature into an actual revenue line. Blocked only on a go/no-go and business verification.
With pilot evidence in hand, invest in a second engineer, a modest hosting/infra tier, and a focused go-to-market motion (schools & learning-support orgs). Retire the bus factor; open the B2B channel.
We have a live, production-grade, accessibility-first learning platform with a built-in revenue engine — created for a fraction of what it would normally cost. The expensive, risky part (building it) is done. What's left is the fun part: putting it in front of students and turning the revenue on.