CrystalClear

Your pocket crystal oracle ♡

Comment

Submitting To

CrystalClear is a submission to the CS Girlies - AI vs. HI hackathon, to the Make Anything, But Make it YOU track.

Inspiration

I've always loved the cozy, magical aesthetic of crystals, and I wanted to create an app that blends that vibe with actual tech functionality. CrystalClear was inspired by the idea of turning everyday crystal curiosity into a personal, spiritual experience, regardless of whether you’re new to crystals or deeply into their meanings.

What it does

CrystalClear lets users take or upload a photo of a crystal, identifies it using the Google Cloud Vision API, and then returns key, identifying traits and spiritual properties of the crystal. Users can save their favorites to a personal "altar", and they can also get AI-generated readings of the crystals they've scanned.

How I built it

I built CrystalClear using React Native and Expo for cross-platform support and media access. The app sends crystal photos to the Google Cloud Vision API to get label predictions, then uses Firebase Firestore to fetch and store crystal data, including traits, metaphysical properties, and saved altar entries.

Challenges I ran into

One major challenge was working around the limitations of label detection. Crystals often look alike, so I tried my best to ensure the output still felt meaningful even when accuracy varied. Integrating multiple services (camera, API calls, Firebase) smoothly within a mobile UX also required careful debugging and testing.

Accomplishments that I'm proud of

I've never taken part in a hackathon solo before, nor have I built a project of this caliber on my own! I'm so proud of what I've created. I'm also proud of how cohesive and cozy the experience feels, from the design to the functionality.

What I learned

I learned how to orchestrate multiple services (like image recognition and real-time databases) in a way that feels smooth and responsive in a mobile context. I've also worked with Expo for the first time, and found it to be incredibly useful. I'll definitely be using it with other projects!

What's next for CrystalClear

I'm hoping to improve crystal detection with custom-trained models, expand the altar feature with journaling, and maybe even introduce daily rituals or notifications based on users’ saved crystals. I'd definitely love to release this for public use at some point!

Built With

Share this project:

Updates