TLDR: I built The Early Days — a voice-first journal for new parents. Tap and talk, get a written, searchable record of the early weeks. On-device speech-to-text for accuracy and privacy. Plus postcard sending via Lob. Built before my daughter Lydia arrived. Currently in active use.

The Spark

Preparing for parenthood meant months of research — gear, delivery planning, nursery setup. But as my due date got closer, something clicked: I was spending all this energy preparing for having a baby, but not for remembering it.

The newborn phase moves at a disorienting pace. Days blur together. You want to remember the small moments — the first smile, the weird sounds at 3 AM, how they fit in your palm. But you’re also exhausted and one-handed and opening a notes app feels like friction you just don’t have energy for.

That’s where The Early Days started: a solution to a personal problem I was about to face.

The Problem with Existing Options

New parents want to document the early weeks. The timing is terrible:

• One-handed (holding a baby)
• Sleep-deprived (not typing clearly)
• Emotional (wanting to capture feeling, not just facts)
• Overwhelmed (no mental bandwidth for complex apps)

Traditional journaling apps require typing. Voice memos are just audio files — unstructured, hard to search, impossible to skim. Photos exist but lack context. There’s a real gap for a tool that lets you speak memories into existence and instantly get a written record.

What I Built

The Early Days is a voice-to-text parenting journal with one core premise: speak your memories, get a written record.

The key features:

• Voice-first interface — tap and talk, no typing required
• Automatic transcription — speech-to-text converts your voice into written entries immediately
• Timestamps and tagging — every entry captures when it happened with tags for searching later
• Searchable journal — find moments by date, keyword, or tag (“first smile,” “3 AM thoughts,” “feeding struggles”)
• One-handed UI — the entire interface is designed around the reality of holding a newborn

The design philosophy is ruthless minimalism. No social features. No analytics. No friction. Speak → record → remember.

Technical Choices

The stack is Next.js and Supabase, and both were chosen for the same reason: simplicity at the cost of almost nothing.

Next.js gives enough flexibility to scale if the app grows beyond personal use, and server-side rendering actually matters here — new parents are often on phones with patchy connections, in hospitals or nurseries, at weird hours. Pages need to load fast.

Supabase handles auth, database, and file storage in one place. For a solo project, that consolidation matters more than any individual feature.

On-Device Speech-to-Text

The speech recognition runs on-device using a local transformer model rather than a cloud API. This was a deliberate choice for two reasons: accuracy and privacy.

Cloud STT services are general-purpose. They’re optimized for clear speech in normal conditions. A tired parent whispering at 2 AM while holding a baby is not normal conditions. Running a local model means I can tune for the use case — quieter audio, informal speech, baby-related vocabulary — without sending audio to a third-party server. The notes you record about your kid are personal. They shouldn’t be processed on someone else’s infrastructure.

The model runs entirely in-browser via transformers.js, which means no network call and no latency waiting for a round-trip. Tap, speak, done.

The data model is as minimal as the UI:

• Entry — timestamp, transcribed text, tags, voice file reference
• User — basic auth via Supabase

Postcard Sending via Lob

One feature I’m especially happy with: you can send a physical postcard from inside the app using Lob. Pick an entry, attach a photo, and Lob handles printing and mailing a real postcard to grandparents, family, whoever.

The idea came from realizing that digital journals are great for you, but grandparents don’t have logins. A physical postcard with a photo and a few sentences from your journal entry is something they can actually hold. It costs about $1.50 to send and takes maybe 30 seconds to set up in the app. It’s one of those features that feels small until someone’s grandmother gets a postcard in the mail with a photo of a baby she hasn’t met yet.

The Emotional Layer

What makes The Early Days different from a generic voice diary is understanding why new parents want to journal. It’s not self-improvement or reflection — it’s that time passes too quickly.

Parents look back at photos from a month earlier and feel genuinely shocked at how much has changed. The first weeks are disorienting and beautiful and hard simultaneously, and you want to hold onto them while you’re in them. This app is built around that impulse.

Most apps try to add features. This one intentionally removes them. In the chaos of new parenthood, a tool that gets out of the way and lets you focus on the moment — not the app — is genuinely valuable.

Current State

The Early Days was built in preparation for my daughter Lydia, and it’s actively in use capturing the early weeks of her life. It’s currently private, but the learnings from building it are clear:

• Parents want to journal but lack the energy for any friction
• Voice input directly solves the one-handed, exhausted problem
• Simplicity isn’t a limitation — it’s the point

The app is meant for a specific season of life. It’s not meant to be used forever. It’s meant to be useful exactly when it matters most, then become a beautiful archive you return to years later.

More to come — including a potential open-source release and any community features that make sense based on early feedback.

If you’re expecting or in the early days of parenthood yourself, check out earlydays.app. And if you have thoughts on what would make it more useful, I’d love to hear from you.