[Draft] Building Morning Intent Alarm

This is a work in progress. I'm documenting as I go — expect updates, rough edges, and occasional tangents.

I've worked on products for startups and teams for years, always for someone else and always within someone else's constraints. I often had shower thoughts and random product ideas while walking or doing something, but I never acted on most of them as they seemed like spontaneous ideas not worth pursuing.

Now that AI tools are changing how quickly ideas can become real, I thought — why not try building something? Document the mess and see what happens. The Morning Intent Alarm might not be the first product I've built, but it will be one where the mess, the process, and the questions in my head are documented as I go.

I started with a simple question: What do I actually want to solve? Not a market gap. Not a business opportunity. Just what's annoying in my own life.

Among the things I brainstormed, one kept coming back: waking up. Or rather, waking up fresh. It's honestly a pain of my existence.

Since this is purely a creative endeavor, I'm skipping most of the sanity check as I'm building this for the fun of it, not for a business pursuit. The focus remains on the core issue: the problem.

Problem Framing

Most alarms feel like an attack. Jarring tones, generic music, motivational speeches from strangers that have nothing to do with your life. You wake up reactive, scattered, already behind.

But the alarm itself isn't really the problem.

The problem is that waking up is a pain for me. Not physically — just mentally. I open my eyes and my brain is off. I know I planned things out the night before. I know there are things that matter. But that part of my brain hasn't booted up yet.

So I do what everyone does: reach for my phone. Doom scroll. Check notifications. Start the day on autopilot.

On intense days, I don't have this issue. If something big is happening, I'm looping through it all night — anxious, hyper-aware. I wake up already in it. But most days aren't like that. Most days, I have a good grasp of what's important and what needs to happen. It's just not accessible in those first few minutes. It takes time before that memory kicks in, before the intention resurfaces.

And by then, I've already wasted the start of my day reacting instead of acting.

Traditional alarms don't help with this. They get me out of sleep, but they don't get me into the day with clarity. I'm awake, but I'm not present.

What I actually need is something that reminds me what I already decided mattered — right when I wake up, before autopilot takes over.

The Idea

The concept is simple: let night-you brief morning-you.

Each night, the app presents a prompt. You record your answer in your own voice — could be 15 seconds, could be 2 minutes, depends on how much energy you have and how far you want to go. That recording becomes your alarm. You wake up hearing yourself, with context already loaded.

It's not motivation. It's not a pep talk. It's just you telling yourself what you already decided mattered.

How It Works

Each night, the app presents a prompt → you record your answer → that becomes your alarm → you wake up hearing yourself, already oriented.

Recording the night before forces a moment of reflection when your brain is still on. You're deciding what matters before you shut down for the night, not scrambling to figure it out in the fog of waking up. And there's something about speaking it out loud — it makes it more concrete than just thinking it. It's a commitment. A verbal contract with yourself.

Your own voice matters because familiarity cuts through the fog faster than a stranger's voice. It's not generic advice. It's your plan, your words, your context. And it's harder to dismiss. When it's you talking to you, there's continuity. Night-you prepared something for morning-you. That's harder to snooze than a random chime or someone else's motivational speech.

Later versions might cycle in older recordings — principles, reminders, motivations from past-you that still matter. But that's a nice-to-have. For now, it's just about tomorrow.

The Prompts

The prompts are short, clear, and designed to feel like a nudge — not homework. Here are a few I'm starting with:

Action-oriented — what to do:

"What's the first thing I'll do when I get up?"
"What's one thing that moves the needle today?"
"If I only get one thing done, what should it be?"
"What's waiting for me that I'm ready to tackle?"

Reflection-oriented — how to think:

"What am I looking forward to?"
"What would make today a win?"
"Who am I showing up for today?"
"What's one thing I don't want to forget?"

The goal isn't to pick the "right" prompt. It's to give yourself a structured way to load your intention into your alarm.

What's Out There

You'd think someone would've solved "better wake-ups" by now. Turns out, there's a lot — but most of it misses the point.

The voice recording apps

Apps like My Voice Alarm ($0.99) let you record your own voice as the alarm sound. You can build a library of recordings and loop them until you wake up.

The problem? Your voice is just another sound. There's no prompt. No ritual. No decision baked into the act of recording. You could record "GET UP!" and it works the same as any loud alarm — it just happens to sound like you.

The motivation apps

Alarmy (75M+ users) makes you solve math problems or shake your phone. Avo uses AI voices of celebrities to deliver custom messages with your calendar, weather, and news. Some apps even have AI call you and hold a conversation until you're awake.

The gap? All the motivation comes from outside. It's someone else's voice. Someone else's words. Even when it's "personalized," it's still generic — stitched together from data, not from you. It doesn't know what you decided mattered last night. It's guessing.

The morning routine apps

Apps like Fabulous and Routinery help you build habits after you're up — guided routines, timers, streaks.

The problem? They assume you're already awake and functional. They don't help with those first foggy minutes when your brain hasn't booted yet and you're reaching for your phone to doom scroll instead of starting your day.

The insight: The market has solutions for waking up and solutions for building routines — but nothing bridges the two. No one is asking: what do you need to hear in that first disoriented minute?

That's where Morning Intent Alarm fits.

Scoping the MVP

The first version is a proof of concept, a demo. The goal isn't to build a full-featured alarm clock yet. It's to test one thing: does hearing my own voice with my own intention actually help me start the day better?

Everything else is noise until I know that works.

What's in v1

The absolute bare minimum:

1. One prompt question
Not a library. Not rotation. Just one prompt that works. I'll probably start with "What's the first thing I'll do when I get up?" — simple, action-oriented, hard to overthink.

2. Voice recording
Hit record. Say your answer. Done. No editing. No fancy audio processing. Just capture the voice memo.

3. Re-record if needed
If you fumble or want to redo it, you can record again. But no library of past recordings. Just replace it.

4. Set the alarm time
Pick what time you want to wake up tomorrow. That's it.

5. The alarm plays your recording
In the morning, the app plays a normal alarm sound for the first 5 seconds to actually wake me up, then transitions into playing my voice recording. Loops until you stop it. There's a stop button to turn it off.

6. Repeat the next night
You can do this again the following night. Same flow. New recording.

That's the entire app.

What's tempting but waiting

Multiple alarms — Not yet. One alarm. One morning. One intention.

Snooze — Maybe later. For now, the recording just loops. If you want to snooze, you stop it and set another alarm manually. Not ideal, but not the point of v1.

Prompt rotation or changing prompts — Tempting, but no. One prompt. I need to know if the core mechanic works before I add variety.

Cycling old recordings (principles, past motivations) — Definitely later. This was in the original idea, but it's not essential to test the hypothesis.

Calendar or task integration (especially with task management apps) — Would be cool. Not now.

Analytics, streaks, or habit tracking — Nope. I'm not trying to gamify this yet. Just trying to see if it works.

Custom alarm sounds, fade-in, vibration patterns — All the normal alarm clock features can wait. This isn't about being a better alarm clock. It's about testing if self-intent works.

Design priorities

Bright colors — I want this to feel energizing, not clinical — something that gets my eyes open
Simple and fast — Minimal taps. No onboarding hell. Open app → record → set time → done.
Beautiful but basic — Clean UI, but no over-design. Function first.
Works offline — This is too tied to routine to depend on internet. Everything runs locally.

Platform

I'm building this as a mobile app using Replit. The main reason: speed. I want to get from idea to working prototype as fast as possible without getting stuck in development setup or tooling.

An alarm app needs to live on your phone. It's too tied to your routine to be anything else. The good news is Replit builds apps that work on both iOS and Android, so even though I'm testing on my iPhone, it's not locked to one platform.

If this works, I can always refine or rebuild. But for now, the goal is to test the hypothesis — not to build production-ready infrastructure.

Writing for AI (The Vibe Coding Brief)

This is where things get interesting. The usual flow of writing PRDs and technical specs goes like this: PRD → review → iterate → design → wireframes → development → iterate some more. It's a good process. It's thorough. It catches edge cases. But it's also overkill for a proof of concept I want to test in a weekend.

Writing requirements for AI is different. Not better or worse — just different.

How it's different from traditional specs

Traditional specs are about eliminating ambiguity. You document every state, every edge case, every pixel. You write user stories. You create acceptance criteria. The goal is to hand off something so clear that a developer (or a team) can build it without having to ask clarifying questions.

AI thrives on vibe and intent. You don't need to be exhaustive. You need to be clear about what matters and flexible about how it gets there. You can describe the feeling you want, the flow you're imagining, the constraints that are non-negotiable — and let the AI fill in the gaps. If something's off, you iterate. Fast.

It's less about pixel-perfect wireframes and more about: here's the problem, here's the core experience, here's what should feel easy, now build me something that works.

The trade-off? It's probably not scalable. But it's fast. And for a proof of concept, fast matters more than perfect.

The approach

For this build, I'm using Replit Agent — an AI coding assistant that can scaffold and build the entire app. I'm not handing it a 10-page spec. I'm giving it context, intent, constraints, and some design direction. Then I'll iterate as it builds.

The prompt is conversational but structured. It includes just enough technical direction to avoid chaos, and just enough flexibility to let the AI make decisions I don't care about yet.

Here's what I'm feeding it:

# Morning Intent Alarm - Build Brief

## Context

I'm building a proof-of-concept mobile alarm app called **Morning Intent Alarm**. The core idea: let night-you brief morning-you. Each night, you record a short voice memo answering a prompt (e.g., "What's the first thing I'll do when I get up?"). That recording becomes your alarm. In the morning, you wake up hearing your own voice with your own intention already loaded.

This is testing one hypothesis: does hearing my own voice with my own intention help me start the day better than a traditional alarm?

Everything else is secondary.

## What to Build

### Core Features (v1 MVP)

**1. Night Flow**
- Open Morning Intent Alarm
- See today's prompt: "What's the first thing I'll do when I get up?"
- Tap to record voice memo (no time limit, but designed for 15 seconds to 2 minutes)
- If you don't like it, re-record (replaces the previous one, no library)
- Set alarm time for tomorrow morning
- Done

**2. Morning Flow**
- At the set time, the alarm goes off
- Plays a standard alarm sound for 5 seconds (to wake you up)
- Transitions to playing your voice recording (loops until stopped)
- Big, obvious stop button
- That's it

**3. Repeat**
- Next night, same flow with Morning Intent Alarm
- New recording replaces the old one

### What's NOT Included

- No multiple alarms
- No snooze functionality (recording just loops)
- No prompt rotation or customization
- No streaks, analytics, or gamification
- No fancy audio editing or effects
- No calendar integration
- Keep it brutally simple

## Design Direction

### Visual Style
- **Bright, energizing colors** — think morning energy, not sleepy blues
- **Clean and minimal** — no clutter, no overwhelming UI
- **Big, tappable elements** — especially the record button and stop alarm button
- **Fast to use** — minimal taps from open → record → set alarm → done

### Tone
- Not clinical. Not corporate. Feels personal and intentional.
- Like a tool you made for yourself, not something designed by a wellness company.
- Morning Intent Alarm should feel energizing but not pushy.

### Key Screens (rough concept)

**Screen 1: Tonight's Prompt**
- Large prompt text at top: "What's the first thing I'll do when I get up?"
- Big record button (center, unmissable)
- If already recorded: show waveform preview + re-record option
- Set alarm time (simple time picker)
- Confirm/Done button

**Screen 2: Alarm Goes Off**
- Full-screen, high-contrast
- Shows the time
- Plays alarm sound → transitions to your recording (looping)
- Big STOP button (can't miss it)

**Optional: Simple home/status screen**
- Shows if alarm is set for tomorrow
- Shows the time
- Quick access to re-record or change time
- Not essential for v1, but nice to have

## Technical Constraints

- **Platform:** Mobile app (iOS primary, but cross-platform is fine)
- **Audio:** Record and playback voice memos locally (no cloud storage needed for v1)
- **Alarm:** Needs to trigger at the set time even if Morning Intent Alarm is closed (use system alarm/notification APIs)
- **Offline-first:** Everything should work without internet
- **Minimal dependencies:** Keep it simple — don't over-engineer

## Non-Negotiables

1. The recording must play even if Morning Intent Alarm is in the background or closed
2. The alarm must actually wake me up (hence the 5-second alert sound first)
3. No onboarding friction — open app and start recording immediately
4. Works offline

## What Success Looks Like

I open Morning Intent Alarm at night. I see the prompt. I record a quick voice memo. I set the time. Done.

In the morning, the alarm wakes me up, I hear my own voice telling me what I decided mattered, and I stop it when I'm ready.

That's it. If that works, this is a success.

Why this structure works

The prompt sets context first — not just what to build, but why it exists. That helps the AI make better decisions when it hits ambiguity.

It's specific about what matters (the core flow, offline support, alarm triggering) and vague about what doesn't (exact colors, button shapes, layout details). I can tweak those as I see what the AI generates.

It uses plain language, not formal requirement syntax. No "As a user, I want to..." or "Given/When/Then" structures. Just: here's what happens, here's what it should feel like, now build it.

And critically, it defines success — not in terms of features shipped, but in terms of the experience working. That gives the AI (and me) a north star to iterate toward.

See It In Action

Now the fun part — handing this off to AI and seeing what comes back.

Fair warning: this isn't a "paste prompt, get perfect app" story. What you'll see at the link is whatever version exists right now — could be the first attempt, could be the tenth. Getting to a working MVP takes iteration. Prompts get refined. Things break. You adjust and try again.

If you want to skip ahead and see where it's at: Morning Intent Alarm in Tinker Studio.