Armando, the former CTO of Amper Music and later Head of AI Audio at Shutterstock, has launched Image to Music AI. The tool does what it says on the tin: upload a photo, get a soundtrack. It runs on Google's Lyria model and analyzes visual cues like color, lighting, and composition to generate matching music. New users get 15 free credits to try it out.
The pitch is simple. Instead of typing music terminology into a prompt box, you just upload a photo. A sunset gets warm tones. A dark, smoky scene gets a moody groove. Generation takes 2-5 minutes, and you can compare multiple versions side by side before downloading. You can also add text prompts to refine results if you want more control.
Armando co-founded Amper Music back when AI composition was barely on anyone's radar. Shutterstock acquired Amper in 2020, and he spent time leading AI audio integration there. Now he's building independently, working with Google's Lyria as the backbone.
Early feedback on Hacker News has been positive. User muzani compared testing the tool to the recent ChatGPT-to-Ghibli trend, saying it produced impressive results even from mundane photos like a picture of hummus. The tool is built for visual thinkers who find it easier to convey mood through images than music jargon. The Amper pedigree makes this worth watching. The hummus test suggests the tech works. The question is whether anyone needs photo-generated soundtracks once the novelty wears off.