Homeword: Language Micro-Lessons

Inspiration

I call my grandparents every week, and staying connected with my family is very important to me. Since they are not fluent in English, I push myself to speak in their language. Over the years, I have improved a lot: I can hold conversations, read, and even text comfortably (or so I like to think!)

I used Duolingo to sharpen my skills, but I kept running into the same problem. When talking about my day-to-day life, like my hobbies, my work, or little details outside of standard lessons, I could not find the right words. Those gaps left me frustrated and often caused my conversations to stall. There were so many important parts of my life I wanted to share, but I simply could not express them because I was missing the words.

What I wanted was a tool that was low touch, that could slip into my daily routine, and help me fill those vocabulary gaps with words that were actually relevant to me.

This matters deeply to me because my family language is the bridge to stronger relationships with my grandparents and relatives. As I go further in my heritage language journey, I want to express myself more fully and communicate my life, not just the basics.

That is also where I saw a gap in the market. Most language apps are designed for beginners, walking through structured lessons. But once you already have a base, it is hard to find tools that help you expand vocabulary in a personalized, meaningful way.

Already using Homeword myself, I have seen the difference! Every day, I get push notification with a terms that fits my life, like "database" for work, along with its translation and an example sentence. Paired with lightweight flashcards, pronunciations, and quizzes, it is a daily nudge that adds real value to my language journey!

What Homeword Offers Today

Homeword is designed to fit seamlessly into your daily routine. Every day, it delivers a personalized micro-lesson directly in a push notification, so you can learn without even opening the app. Each word comes with its translation and an example sentence, helping you grow your vocabulary in context.

Inside the app, Homeword automatically creates flashcards, pronunciations, and quizzes around your words. This makes it easy to practice and review new vocabulary. Lessons are personalized to your interests, profession, and studies, so the words you learn are immediately relevant to your life.

At launch, Homeword supports more than 15 languages, ranging from widely spoken ones like Spanish and Japanese to others that help people reconnect with their family languages. With streak tracking and simple review tools, Homeword helps you stay consistent and keep building your vocabulary over time.

How I Built It

Homeword was built as a lightweight iOS app using React Native with Expo (Expo Router), with all code organized in a monorepo. The backend runs on Supabase, providing authentication, data storage, and lightweight Edge Functions written in Deno, which the app calls directly.

Push notifications are powered by the OneSignal SDK with iOS configuration, delivering the daily micro-lessons directly to the lock screen. For in-app purchases and subscriptions, I integrated RevenueCat. The build pipeline runs through EAS for iOS, with GitHub Actions handling checks, migrations, and staging-to-production promotion.

From an architecture perspective, I kept the flow intentionally simple. Supabase reads and writes are called directly, without queues or a fanout. This let me move quickly while still leaving room for future scaling with patterns like the outbox pattern. I used the Supabase local stack and Expo Dev Client for rapid iteration, which made it possible to test new features such as micro-lessons, push notifications, flashcards, quizzes, and streaks quickly.

Challenges I Ran Into

One of the biggest challenges was the onboarding flow. Because Homeword is personalized, I had to design a process that introduces users to the app, reassures them that I can help with their vocabulary gaps, and demonstrates value right away. Striking the right balance, showing sample push notifications, quizzes, and flashcards without overwhelming new users, took more time than I expected, but it was critical for building trust.

Authentication and external integrations also presented hurdles. I experimented with Twilio for messaging, but the approval process was cumbersome. In hindsight, this friction was a blessing. It forced me to cut nonessential work and focus on the core product experience instead of hyper-fixating on features that did not directly solve the problem.

Finally, I had to make trade-offs in the backend architecture. I initially wanted to build with an outbox pattern running on Kubernetes, but speed mattered more than elegance. For launch, I kept things simple with Supabase core features while keeping a path open for future scaling with tools and patterns.

Accomplishments That I'm Proud Of

One of the biggest accomplishments was personal growth. From the very beginning, I pushed myself to share progress publicly on TikTok, post updates, and engage with potential customers before launch. That was new territory for me, and I am proud of how much more comfortable I have become with putting my work out into the world.

I am also proud of launching a full-stack product under a tight deadline. Building both the iOS app and the Supabase backend, integrating payments, push notifications, and shipping everything before the Shipaton deadline required making hard trade-offs. Instead of chasing every technical ideal, I focused on what mattered most: getting Homeword into the hands of users.

Other accomplishments include clarifying the audience and positioning of the app, designing an onboarding flow that builds user trust, and delivering a polished end-to-end experience that went all the way through App Store review and subscription setup. All of these made Homeword a real, working product.

What I Learned

Building and releasing Homeword has already taught us a lot, both about my users and about the process of building a product like this.

Insights about users

Language learning is not a one size fits all. Homeword has only just been released, yet many of my assumptions are already being tested.

• First and second generation users may be overestimating their fluency. Those who are most engaged with lessons so far tend to already be more fluent, and surveys suggest they use their family or secondary language in multiple contexts such as travel, work, family, or simply for personal joy.
  
• Re-engaging first and second generation users who have lost touch with their family language is likely to be my biggest challenge. This group represents more than 80 million Americans, and whether I can help them reconnect will determine the long term success of Homeword.
  
• Partners are an unexpected audience. More users than expected are motivated partners of first and second generation Americans. Their drive to learn comes from cultural appreciation and a desire to strengthen relationships by engaging with their partner’s family language.
  

Insights about building

I also learned valuable lessons about building the app itself.

• Push notifications are powerful but complex. Delivering meaningful content directly to the lock screen required careful configuration, but it created real user value by making the lessons unavoidable and effortless.
  
• Simplicity wins at launch. Cutting back on more complex backend patterns such as Kubernetes outbox workflows allowed us to move faster with Supabase reads and writes.
  
• Onboarding is everything. A polished onboarding flow matters more than expected, since it is the user’s first proof that Homeword can deliver real value.
  

Surprises

• I expected most of my early users to be motivated second generation learners, but the discovery of partners as a serious audience segment was surprising and opened up new possibilities for growth.
  
• I underestimated how difficult it would be to gauge the right level for users. Some learners benefit from easier micro-lessons as a refresher, while others are ready for more advanced terms. Finding the right balance between challenge and accessibility is an ongoing difficulty.

What’s Next for Homeword

In the short term, my focus is on improving the core product and expanding its reach. Planned features include:

• Adding in-app feedback so users can guide how lessons evolve
  
• Iterating on micro-lessons and running A/B tests to refine what appears in push notifications
  
• Improving flashcards, quizzes, and review flows to make learning smoother
  
• Introducing a free “Word of the Day” version of the product
  
• Expanding beyond the current 15+ languages
  

Longer term, the vision for Homeword is much bigger. I want to become the go-to app for rebuilding the connection with family and heritage languages. For first and second generation populations around the world, Homeword aims to close the vocabulary gaps that make conversations with loved ones harder than they should be, and to help millions reconnect more fully with their culture and identity.