Mixtral AI

• 

  Mixing application

• 

  Image combining Abstract art and Nature remixed with the app

• 

• 

Got you — here’s a clean, concise, Devpost-friendly version. Same impact, way tighter.

Inspiration

AI visuals today are passive—you type prompts and wait. Inspired by how DJs perform music, we wanted to make AI visuals something you can play, feel, and control in real time using physical hardware instead of words.

What it does

Mixtral AI turns AI image generation into a live performance. Using a DJ controller, creators blend images, styles, and moods in real time through knobs and faders—treating AI like an instrument, not a tool.

How we built it

I connected a MIDI DJ controller to an AI-powered visual engine. Physical inputs are mapped to semantic and latent parameters, enabling smooth, real-time blending of images through a web-based interface.

Challenges we ran into

AI models aren’t designed for live performance. Managing latency, smoothing continuous controller input, and maintaining visual coherence during real-time blending were our biggest challenges.

Accomplishments that we're proud of

• Built a fully controller-driven AI visual system
• Enabled real-time AI image blending through physical gestures
• Delivered a live, expressive demo without relying on text prompts

What we learned

AI becomes more creative when it’s performed, not prompted. Physical interaction unlocks intuition and flow that text-based interfaces can’t provide.

What's next for Mixtral AI

We plan to add video, generative motion, multi-layer visual decks, and collaborative live performances for concerts and immersive art.

Built With

• bun
• easymidi
• fal-ai
• next.js
• node.js
• react
• zustand