Wing Date Night - Couples App

AI concierge for couples: Wing learns your relationship—preferences, schedule, quirks—and creates personalized date plans, nudges, and coaching to help you stay close through life’s ups and downs.

Comment

Inspiration

Back in 2018-2020, I ran a surprise date night concierge service exclusively for existing couples. I’d get to know each person in the relationship personally, learn what they loved (and didn’t), and then planned their date nights—delivered step by step as a surprise.

Sometimes that meant planning a four-year anniversary with a budget of $75 (shout out to the Santa Monica Playhouse for making this a special night!), and other times it meant organizing a $2,000 experience with a Michelin-star dinner and a private boat for watching fireworks with the most precious golden doodle

It was incredibly manual. I hacked together surveys, Airtable workflows, and spreadsheets to match each couple’s preferences with hidden gems in their city.

I'd call up small venues and restaurants to ask for special touches in the experience:

"Mr. Smith, could I offer you a glass of Glenlivit 18?", (client's favorite drink)

"A couple I'm working with is on a post-escape-room scavenger hunt, do you mind hiding a pastry coupon somewhere in your cafe?"

My dream was to build a "Stitch Fix for Date Night"—a service that would learn your preferences and deliver curated, delightful experiences in real life. But I couldn’t figure out how to scale it or monetize it, and eventually COVID pulled the plug.

Still, the problem space stuck with me. There are countless apps that help people get into relationships—but almost none that help couples stay connected, stay curious, and stay spontaneous once they’re already together. That’s what Wing is here to change.

What it does

Wing helps busy couples plan thoughtful, personalized dates that feel fresh and exciting. Users answer a few questions about their relationship, preferences, and mood, and the app uses LLMs to generate full, multi-step date itineraries.

It’s not just about where to go. Wing also includes an AI-avatar relationship coach, dedicated to helping talk through difficult topics: money challenges, life stages, household duties, love languages, and more.

The app shares conversation starters, thoughtful prompts, and fun little nudges to help couples connect more deeply. And it keeps things playful too—like finding out Wannabe by the Spice Girls dropped on the same day as your anniversary.

How we built it

The app is built from scratch using Bolt.new, which made it easy to go from idea to execution Supabase handles auth, the database, edge functions, and real-time interactions. OpenAI and Perplexity power the LLM logic that creates each personalized experience, with plans for tool calling for weather forecast and local events / availability. When I got stuck on difficult backend tasks, I worked with a number of AI IDEs / plugins to help, one even went rogue and temporarily went haywire in my supabase instance.

Tavus handles the relationship coach, RevenueCat manages the paywall, which kicks in after three free itinerary requests, Google Maps Places API manages location autocomplete.

Challenges we ran into

There were quite a few. The big challenge was the constant evolution of the idea. I started with a rough “phase 1” plan, but not much beyond that. As I worked on it, the UI, the data model, and even the feature set kept shifting. That meant I had to keep revisiting earlier decisions, which slowed things down.

Specifically, Auth and a lot of the front-end/back-end handoffs turned out to be way more complicated than expected. Looking back, I probably would’ve left Auth out entirely until the core features were solid. As I was testing out some other AI tools, one decided to manipulate my database schema and undoing that was about 5 hours of stress.

Accomplishments that we're proud of I honestly didn’t think I’d get this far in such a short amount of time. This is an idea I’ve been carrying around for years, and finally sitting down to build it felt like cracking open a time capsule.

One of the coolest moments came while brainstorming with Bolt. It suggested an idea I hadn’t considered—showing fun facts and historical moments tied to important dates in a couple’s timeline.

Wing might let you know: – 🎵 A famous song was released on this exact day – 💍 Your 1,000th day together is coming up – 🌟 You share an anniversary with Beyoncé and Jay-Z

It’s a small detail, but those unexpected touches add a lot of charm.

What we learned

Two big lessons stand out.

First, I realized it’s never been faster to go from an idea to something real, dare I say, production. Back in 2018, building a service like this was janky, comprised of form builders, one-on-one interviews and spreadsheets of cool things that I had curated in the city that I was in. I realize now that what I wanted to build then simply wasn't possible.

Second, I continue to learn that planning bite-sized chunks of work while keeping the end-goal in mind remains the best way to get a good result out of LLMs. The LLM needs context, and if it fails to fix or do generally what you had wanted, chances are you need to revert and provide more context + "ask me if you have any clarifying questions" instead of "no, wrong, pls fix" a dozen times.

What's next for Wing

This challenge has been incredibly inspiring, and encouraged me to take action. The immediate goal is to get Wing into the App Store with the core feature set live and running on a simple subscription model (e.g. monthly subscription, includes both partners).

  • Coaching that helps with real stuff: support couples through trickier topics like money, long-term planning, and staying connected.

Progressive learning: Instead of dumping every question up front, I want Wing to keep learning about the relationship over time, naturally, so it builds a better understanding of both people. I've tried implementing "cooldowns" on connection prompts so it's not always in the UI, but there's work to be done.

More personalized moments: milestone reminders, holiday-specific suggestions, gift planning, and nudges that actually feel thoughtful (not generic).

Calendar integration: So couples can plan dates that actually fit into their lives.

Deeper insights: Right now, it focuses mostly on joint interests. I’d love to bring in more about each individual person and what they’re into, what they care about.

Better prompting and tooling: Fast follows include improving how dates are presented (summary + details), partner invites, deeplinks, notifications, and function calling for stuff like weather or local events.

Design polish / consistent styles: Always room to level this up visually.

Built With

Share this project:

Updates