Mentoré

A personal AI coaching platform where coaches understand you deeply—because they're powered by your own journaling, life tracking, and accumulated personal context.

Comment

Mentoré - Shipyard Hackathon

TL;DR

Mentoré is an AI coaching platform that turns journaling and life tracking into personalized guidance.

  • Coaching first, not productivity admin.
  • Personal context over generic chatbot replies.
  • Web + iOS + Android from one monorepo.
  • Instant setup with minimal friction, so users can start in seconds.

Inspiration

I built Mentoré because I wanted a better way to understand what actually gives me energy and what drains it.

Most tools made me do too much setup and gave too little insight. I wanted to journal and track my life without overthinking it, then zoom out and see real patterns: weeks where exercise helped, habits that improved focus, or relationships that quietly pulled me off course.

For me Mentoré is built around small experiments:

  • Try waking up at 6am for two weeks.
  • Move your body more consistently.
  • Then use your coach to reflect on what actually changed and decide what to do next.

For someone else, Mentoré can look different:

  • Career clarity: log weekly wins, missed opportunities, and decision points, then work with your career coach on where to focus next.
  • In-depth training: track sessions, intensity, recovery quality, and consistency over time, then use a fitness coach to identify what is actually improving performance.
  • Decision-making support: even without tracking anything formal, you can get grounded guidance from your coach when you need it the most.

The core idea is simple: a coach should not just collect personal context. A coach should help you interpret it and decide what to do next.

About Me (Builder Story)

I am a full-stack developer with 5+ years of professional experience building scalable, high-performance systems across enterprise and agency environments. I currently contribute to large-scale retail platforms with significant traffic and performance demands.

Before that, I often shipped end-to-end products as the sole developer, owning architecture and technical direction across React, Next.js, Vue/Nuxt, TypeScript, and Supabase.

My background blends engineering and human service. Before focusing fully on software, I founded and operated two restaurants and led teams through rapid growth and COVID disruption. That shaped how I design systems: they only work if they genuinely serve the people using them.

The Problem

People are overloaded with data but under-supported in making sense of it.

Current tools have three big gaps:

  • Admin burden: too much setup, too little value.
  • Passive journaling: entries are stored but rarely synthesized.
  • Context gap in AI: most chatbots do not know your history, patterns, or signals.

If AI cannot remember meaningful personal context, it cannot give meaningful guidance.

What Mentoré Does

Mentoré is a personal AI coaching platform with a Personal Context Engine.

Every journal entry, tracking signal, and chat interaction feeds long-term memory. When you ask for guidance, Mentoré retrieves relevant context from prior weeks or months and responds with grounded coaching.

Put simply:

  • Better personal context leads to better guidance.
  • The more you use Mentoré, the more relevant and useful it becomes.

Core principles

  • Coaching first: reflective guidance over task nagging.
  • Low-friction start: minimal to no onboarding (3 skippable quick questions), then immediate use.
  • Rich context without entry burden: users can ask the coach to create journal entries and auto-track items (for example, meditation minutes) directly in chat.
  • Flexible input style: journal and tracking pages are still available, so each user can choose more structure or less structure based on what works for them.
  • Compounding context: the more you use it, the more useful it gets.

How We Built It

Mentoré is built as a production-oriented cross-platform system:

  • Monorepo: Turborepo + pnpm.
  • Web: Next.js 16 + React 19 + Tailwind CSS 4.
  • Mobile: Expo + React Native (iOS and Android).
  • Shared contracts: TypeScript + Zod (@repo/schemas) + shared business logic (@repo/shared).
  • AI layer: Vercel AI SDK with OpenAI and Anthropic.
  • Data layer: Supabase (PostgreSQL + pgvector + auth + RLS).

Context loop

  1. Input: journal or tracking signal.
  2. Processing: embedding + storage.
  3. Retrieval: relevant personal context fetched with RAG.
  4. Insight: coach responds with grounded guidance.
  5. Repeat: memory quality improves over time.

Challenges We Ran Into

  • Building across web + backend + mobile in ~3 weeks: I started a bit late and prioritized building web first for faster iteration, then expanded to full platform scope.
  • Cross-platform parity: keeping web and mobile aligned in both features and UX while both are evolving quickly.
  • RAG quality at personal scale: turning unstructured journals, chats, and tracking into consistently high-quality retrieval has worked well so far, and we are continuing to optimize precision and relevance through deeper testing.
  • Payments + production readiness: RevenueCat/mobile setup, plus all the practical app-store and release pipeline complexity even for test builds.

Accomplishments We Are Proud Of

  • Working cross-platform product: web + native mobile apps are running from one monorepo.
  • Personal Context Engine is live: meaningful retrieval is working and already improves coaching quality.
  • Low-friction onboarding: lightweight entry flow (3 quick questions), then immediate coaching access.
  • Monetization foundations are in place: tiered pricing and payment infrastructure established.

What We Learned

  • Guidance beats management: reflective coaching feels more human and more sustainable than rigid tasking.
  • Context makes AI useful: memory transforms coaching from generic replies to relevant, personal guidance.
  • Shipping cross-platform quickly is possible: but only with strict shared architecture and strong prioritization.
  • There is still more validation to do: we are proud of the foundation and focused on deeper testing and feedback next.

What Is Next

Near term

  • Bring web and mobile into tighter feature parity.
  • Finalize unified subscription/payment structure across platforms.
  • Expand real-world testing and feedback loops for coaching quality.
  • Prepare full production launch paths.

Future roadmap

  • Voice-first journaling and coaching workflows.
  • Deeper integrations (calendar, health signals, and other context sources).
  • Better proactive insights and pattern detection.

Why This Matters

Most apps help you collect information. Mentoré helps you reflect on your life and act on it.

That is the difference between tracking data and building clarity.

Testing access note: We have not publicly shared iOS/Android testing links yet due to AI usage costs. Once the apps are live in the stores, links will be shared. In the meantime, you can view the website and register interest here: https://www.mentore-ai.com/signup

Built With

Share this project:

Updates