Home / Case Studies / Sounditout

Sounditout

An AI-driven mobile application that bridges the gap between visual memories and music — using computer vision and GPT to generate personalised song recommendations from photographs.

Client
Sounditout
Status
In Progress
Services
AI · Mobile Development · UX Design
Location
Australia
Sounditout — music and AI

Project Overview

Sounditout is an innovative social media application that brings together artificial intelligence, music discovery, and personal memory. The concept is elegant: upload a photo, and the app analyses the mood and emotional tone of the image to surface five songs that perfectly capture how it feels.

Kurrajong Studio partnered with the Sounditout team to design and build the full application — from the multi-layered AI architecture to the cross-platform mobile experience and social features that let users share their sonic memories with the world.

The Challenge

  • Analysing the emotional tone of a photograph accurately enough to make musically meaningful recommendations requires multiple AI models working in concert.
  • Building a real-time pipeline that processes image uploads, runs computer vision analysis, generates semantic embeddings, and returns ranked song recommendations — fast enough to feel responsive.
  • Designing a cross-platform mobile experience (iOS and Android) that feels native, social, and delightful while managing a complex AI backend.
  • Creating a sustainable freemium model that balances user accessibility with monetisation through tokens, ads, and premium features.
  • Building social discovery features on top of a personalised AI engine without compromising the intimate, personal nature of the experience.

The Solution

We built a multi-layered AI architecture with a cross-platform mobile front end — designed to feel effortless despite the complexity running beneath the surface.

How It Works

1
Upload — User uploads a personal photo from their camera roll.
2
Vision Analysis — Azure Computer Vision analyses the image — detecting scenes, objects, colours, and emotional context.
3
Semantic Matching — GPT interprets the visual analysis to extract mood and sentiment descriptors.
4
Song Recommendations — Five songs that best convey the mood and sentiment of the photo are returned.
5
Personalised Playlist — Songs are saved to the user's library, building a musical story of their memories over time.

Cross-Platform Mobile App

Built with React Native and Expo Go for iOS and Android, with Redux managing global state. Rapid iteration and testing without platform-specific divergence.

AI Integration Layer

Azure Computer Vision handles image recognition and scene classification. GPT processes structured output to produce nuanced mood descriptors that drive music matching.

Backend & Database

Java-powered backend with PostgreSQL managing user data, photo libraries, song databases, token balances, and social connections.

Freemium Token System

5 free matches/month, up to 10 additional tokens earned via ads, and premium tiers unlocking genre-based timelines, custom albums, and advanced curation.

Current Status

Sounditout is currently in active development. The core AI pipeline — image analysis, mood extraction, and song recommendation — is operational. The team is iterating on the social features and preparing for initial beta testing with a select user group.

The project represents a genuinely novel intersection of AI, music, and personal memory — and we're proud to be building it alongside a client who had the vision to see the emotional potential of combining computer vision with music discovery.

Tech Stack

JavaReact NativeExpo GoReduxPostgreSQLGPTAzure Computer VisioniOSAndroid

Services Delivered

  • AI Integration
  • Mobile Development
  • UX Design
  • Backend Engineering
  • Database Architecture
  • Product Strategy

Have an AI product idea?

We'd love to help you bring it to life.

Get in Touch