Decoding Music with AI: Built with React, Astro, and Fine-Tuned LLMs
I built Describe Music to solve my own frustration with manually analyzing music. It's an AI tool that instantly breaks down any track's genre, mood, BPM, key, and instrumentation.
The stack is modern and intentional:
Frontend: React + Astro for a blazing-fast, SEO-friendly experience.
AI Core: Custom fine-tuned Large Language Models (LLMs) for audio understanding, not just simple classifiers.
Principle: Privacy by default. Your uploads are processed and immediately discarded—no storage, no data harvesting.
We're at a cool inflection point: Over 2,300 creators are actively using it, performing ~4.2 analyses per session. The feedback on accuracy (especially on obscure genres) has been fantastic.
I'm here for your unparalleled insight:
For Users: Does the analysis depth actually improve your workflow? Where does it fail?
For Builders: How would you leverage this? Is a real-time API the most valuable next step?
For Marketers: Where do music creators truly hang out and discover tools online?
Try it and share your thoughts: https://describemusic.net/
Your feedback will directly shape the roadmap. Thanks for being an amazing soundboard!
This is really smart! I've worked with a few music-related clients as a developer, and they always struggled with manual music cataloging. How are you handling the fine-tuning for obscure genres?
This is really cool. Nailing obscure genres is something most tools miss, so if yours is strong there that’s a big edge. A real-time API could open up a lot of creative use cases—thinking DJs or live-stream setups where instant analysis matters.
Truly appreciate you highlighting our focus on niche genres – that validation means a lot! We're currently hyper-focused on refining accuracy for obscure tracks before expanding capabilities. If you have time, I'd be deeply grateful if you could test it with your most unconventional work and share raw feedback on where the analysis shines or falls short. Your expertise would directly shape our priorities.
Nice Work 👍🏻
Thank you for your encouragement