Orbora

Personalized audio sessions on demand. Generate talks for your mood across meditation, ASMR, and breathwork, or follow guided paths with an animated companion orb

Comment

Inspiration

Generic meditation and wellness apps often feel static and impersonal. We listen to the same pre-recorded voices repeating the same scripts, regardless of what we are actually going through in that exact moment.

With recent breakthroughs in Generative AI, especially natural, emotionally nuanced, human-like speech synthesis—I realized audio talks could be personal, adaptive, and intimately close to the listener.

I envisioned Orbora: a mindfulness companion that creates bespoke guided meditations, breathwork, ASMR, and wisdom monologues on demand. To anchor this experience visually, I created the Orb an omnipresent living visual companion whose breath, color glow, and orbiting rings reflect your emotions, topics, and duration, creating a sense of calm and presence across the entire app.

What It Does

Orbora lets users generate personalized audio talks in seconds:

  • Two Generation Paths: A peaceful Guided Flow (step-by-step questions where you hold the glowing Orb to generate) or a Quick Form for instant generation.
  • AI Audio Engine: Gemini writes the script tailored to your context and intent, and Gemini TTS voices it with expressive, human-quality audio.
  • Dual Library: A Wellness Journey featuring structured, progressive paths (Breath, ASMR, Wisdom, Meditation) with unlockable milestones, alongside a Personal Journey diary of all your generated sessions.
  • Offline-First Playback: Talks and audio are cached locally with resume-playback position memory, 10s seek controls, an expandable full-screen player, and an adaptive docked mini-player.
  • Fair & Sustainable Monetization: Powered by RevenueCat, featuring virtual currency (TCR Talk Credits) and the AuraOrb Pro subscription.

How I Built It

  • Kotlin Multiplatform (KMP) & Compose Multiplatform: A single shared codebase powers Android, iOS, and Wasm verification targets with strict Clean Architecture and MVVM.
  • Native Audio Engines: Leveraged platform-native audio players (Media3 / ExoPlayer on Android and native media frameworks on iOS) with lock-screen controls and background audio playback.
  • Backend on Google Cloud Run & Ktor: A lightweight, high-performance Ktor backend coordinating Vertex AI Gemini, Gemini TTS, Cloud Firestore, and Google Cloud Storage.
  • Flexible In-App Purchases: Integrated subscriptions and consumable credit packs through RevenueCat, paired with custom credit calculation and server-side webhook reconciliation.
  • Local Persistence: Room 3 KMP with bundled SQLite driver for structured data and DataStore for persistent preferences.

Challenges I Ran Into

As a solo developer, building and shipping this for the Ship-a-ton pushed me into unfamiliar territory:

  1. First-Time iOS Development: This was my first time releasing an iOS app. Bridging background audio playback, system interruptions, lock-screen media controls, and platform lifecycles with the shared multiplatform architecture was a steep learning curve.
  2. First-Time Subscription & In-App Purchases: Navigating store configurations for the first time required careful design. I had to cleanly separate sandbox testing from live production environments so internal testing could run smoothly without storefront rejections.
  3. Protecting the Backend & Managing AI Costs: Generative AI audio has real compute costs. To make sure the app remains financially sustainable without losing money on generation, I:
    • Designed a clear unit-cost rule: ( \text{Cost} = 1 \text{ credit per started 30 seconds} ).
    • Secured backend endpoints with cryptographic token verification and App Check.
    • Added weekly usage caps and SHA-256 email fingerprinting to prevent alias abuse on free credits.
    • Set up automated webhook listeners to instantly reconcile credits on refunds or cancellations.
  4. Handling Background App Exits During Generation: Since generating audio can take a little time, users might start a talk and switch apps. I had to ensure the backend generation and queue state kept running reliably without losing progress or charging the user twice.
  5. Balancing Onboarding: I wanted an onboarding tour that clearly taught the Orb mechanics and navigation without feeling heavy or getting in the way of listening.
  6. Focusing on the MVP: It was tempting to get lost in minor details or secondary features. Strictly prioritizing what mattered most was essential to shipping a rock-solid product on time.
  7. GDPR & Privacy Compliance: Built full account deletion that cascades across databases and cloud audio storage, backed by transparent privacy policies and data protection standards.

Accomplishments That I'm Proud Of

  • Shipping to Both Stores: Taking an idea from zero to a live, production-grade app submitted to both Google Play (176+ countries) and the Apple App Store in time for the deadline.
  • The Living Orb Companion: Crafting a custom-rendered visual identity that feels alive—breathing, glowing, and rotating in harmony with the session's emotional tone and length.
  • Sustainable Multiplatform Monetization: Bringing a smooth subscription and virtual currency experience to life across both mobile platforms.

What I Learned

  • How to architect and deliver a production-grade Kotlin Multiplatform app spanning Android and iOS.
  • The intricacies of mobile monetization, store sandboxes, paywall flows, and webhook validation.
  • How to model generative AI economics sustainably within consumer mobile apps.
  • The discipline of cutting non-essential scope to deliver a cohesive MVP.

What's Next for Orbora

  • Quick Anxiety Relief: A dedicated, one-tap instant grounding session for moments of acute stress or panic.
  • Post-Session Conversations: An interactive, conversational reflection where you can talk with the companion about what was said, how you felt, and what you took away from the session.
  • Multilingual Audio Generation: Expanding voice synthesis so users can generate personalized talks in multiple languages.
  • Smart Completion Notifications: Background push alerts notifying listeners the moment their custom talk is ready.
  • Curated Breathing Path: Releasing the dedicated 10-step breathwork journey with newly mastered audio.

Built With

Share this project:

Updates

Submission history