Elqra

Elqra is the AI-powered momentum engine that bridges the gap between "I want this life" and "I did something about it today," using multimodal AI to turn big dreams into micro-actions.

Comment

Elqra

Inspiration

We noticed a massive disconnect in the modern ambitious generation. People consume endless content about "dream lives"—travel, entrepreneurship, fitness—but rarely take the first step. The problem isn't a lack of desire; it's a lack of clarity and permission. We wanted to build a tool that doesn't just list tasks, but actively dismantles the fear of starting by making the first step so small and visually compelling that it feels impossible to fail.

What it does

Elqra acts as a "Momentum Architect."

  1. Onboarding: Users input a vague dream ("I want to open a coffee shop").
  2. Plan Generation: Gemini 3 Flash breaks this down into a 5-7 day "Momentum Plan" tailored to their pace (Gentle, Steady, Bold).
  3. Daily Focus: The dashboard hides the clutter and shows only "Today's Bold Move."
  4. Visual Verification: Users snap a photo of their work. Gemini Vision analyzes the image to verify completion and award XP.
  5. Dream Visualization: We use Veo 3.1 to generate a cinematic video of the user's future success, creating emotional buy-in.
  6. Grounding: The "Location Scout" uses Google Maps Grounding to find real-world places nearby (cafes, gyms, co-working spaces) where the user can work on their dream.
  7. Mindset Coach: A Gemini Live voice assistant helps users navigate imposter syndrome in real-time.
  8. Monetization (New): We integrated RevenueCat to create a seamless "Pro" subscription model, unlocking exclusive AI models and unlimited video generations.

How we built it

  • Frontend: Built with React 19 and Tailwind CSS for a fluid experience.
  • Mobile Engine: We wrapped the experience in Expo (React Native) using a WebView bridge, allowing us to deploy a native Android APK while keeping a unified codebase.
  • Monetization: We implemented RevenueCat using a hybrid approach: The Native SDK handles the heavy lifting (store transactions, paywalls) on the mobile layer, communicating with the Web layer via a custom postMessage bridge to instantly unlock features upon purchase.
  • Core Logic: We used Gemini 3 Flash for generating the structured JSON action plans due to its high reasoning capabilities and speed.
  • Video: We integrated Veo 3.1 Fast Preview to generate 1080p aspirational videos on the fly.
  • Vision: Gemini 3 Flash is used to analyze user-uploaded photos for "Proof of Work" verification.
  • Voice: We implemented the Gemini Multimodal Live API over WebSockets to create a low-latency, interruptible voice coach ("Elqra Coach").
  • Grounding: We utilized Gemini 2.5 Flash with Google Maps Tool to fetch real-world location data.
  • Director Mode: A custom automation engine written in TypeScript that simulates user interaction for high-quality demo presentations. Pro Tip: We even built a "Mobile Device Frame" toggle so we could record perfect smartphone-style demos directly from the web browser!

Challenges we ran into

  • Hybrid Bridge Communication: Connecting the native RevenueCat SDK to the web-based React UI was tricky. We had to build a robust two-way messaging system so that when a user clicks "Subscribe" on the web, it triggers the native OS paywall, and the success result instantly unlocks the web UI without a reload.
  • Video Generation Latency: Veo is powerful but takes time. We had to implement a robust polling mechanism with user feedback loops (loading states) to keep the user engaged while the video rendered.
  • Browser Security: Implementing "Director Mode" (which auto-moves the mouse and clicks elements) required complex calculation of DOM element positions.

Accomplishments that we're proud of

  • RevenueCat "Bridge" Architecture: We successfully combined the flexibility of the Web with the power of Native Monetization. It feels seamless.
  • "Director Mode": We built a self-driving demo mode that narrates the app's story automatically. It's a unique way to pitch the product without manual intervention.
  • Deep Integration: We aren't just using one model; we are using the entire Gemini ecosystem (Text, Vision, Video, Audio/Live, and Maps Grounding) in a single cohesive UI.
  • Mobile MVP: We went from a web-only concept to a fully installable Android APK with working In-App Purchases in under 24 hours.

What we learned

We learned that Multimodal AI allows for a fundamental shift in UI design. We don't need complex forms anymore; a user can just talk (Live API) or show a picture (Vision), and the app understands. We also learned that Monetization is a feature, not an afterthought—integrating RevenueCat early influenced how we designed our user tiers and value proposition.

What's next for Elqra

  • Calendar Integration: Syncing "Bold Moves" directly to Google Calendar.
  • Multiplayer Mode: "Dream Squads" where friends can share their generated Veo videos and track streaks together.
  • Native Widgets: Building dedicated Home Screen widgets for "Daily Motivation" using our existing Expo architecture.

Built With

Share this project:

Updates