This is a submission for the Google AI Studio Multimodal Challenge
What I Built
I’ve always loved listening to background music while working. It helps me stay productive, focused, and in the zone. But I often found myself jumping between endless YouTube “study beats” or “relaxed mind” and other such channels, none of which really adapted to my task, mood, or environment.
That’s why I built CalmaBeats — an AI-powered app that generates continuous, frequency-tuned background music tailored to whatever you’re working on. It’s a smarter, more personal alternative to generic playlists.
The app takes a brief description of the task, images of the user’s office, room, or environment (uploaded or captured on the app using the camera), and the intended session duration. Using Gemini 2.5 Flash, the app analyzes the vibe of your room or environment— colors, lighting, and style — and combines it with your task description. Then, with Lyria RealTime, it turns that context into seamless, adaptive concentration beats.
The result: Music that feels personal, context-aware, and scientifically tuned to support productivity, whether you need calming alpha waves for reading, low-beta focus for coding, or gamma-inspired energy for deep problem-solving.
How to Use CalmaBeats
- Describe your task: Type a short sentence about what you’re working on. Example: “Designing a pitch deck for a fashion brand.”
- Add your workspace vibe (optional): Upload a photo of your office, room, or the environment, OR use your camera to capture your current setup.
- Set your focus duration: Enter how many minutes you plan to work or study.
- Generate your music: Click Send. CalmaBeats will instantly create a continuous background track tuned for focus and concentration.
- Refine if needed: Not feeling the first result? Tap Shuffle to get a fresh variation.
- Save your session: Save your current task + image, + music settings. Next time, you can reload it from Past Sessions without starting from scratch.
Demo
https://calmabeats-59621488053.us-west1.run.app/
How I Used Google AI Studio
I used Google AI Studio’s Applet creator to build on top of the PromptDJ reference. Inside the app:
• Gemini 2.5 Flash powers multimodal vision, understanding the physical space or room where the task is to take place, extracting dominant colors, mood, styles, and context from the workspace image.
• Lyria RealTime handles continuous streaming of music, with loop-safe generation and real-time steering via WebSocket.
The app runs fully on Google Cloud Run.
Multimodal Features
• Text + Image → Audio: Task description + room/desk photo → AI-curated music prompt → continuous instrumental music.
• Frequency-adaptive logic: Task type automatically maps to alpha, low-beta, or gamma “feels” for optimal focus.
• Vision-driven mood: Room photos guide the emotional tone of the music.
• Responsive, simple music visualizer.
• Session timer: Music length matches the intended task duration, with a countdown visible in the UI.
Top comments (0)