Cyanite Senior Research Engineer Benno Weck co-authored two benchmarks investigating how well LLMs actually understand music.

Cyanite gives music companies, and the AI agents increasingly built into their workflows, a deep, structured understanding of music, powering search, recommendation, metadata, and AI-powered workflows at scale. Our technology analyzes the audio itself and makes that understanding available through our Web App and API, or directly within existing products and workflows. At Cyanite, we help our partners: 🎯 Tag music at scale across genre, mood, energy, instruments, vocals, tempo, use cases, lyrics & more 🔍 Search and discover music using natural language and sound-based similarity 🤖 Build human and agentic workflows on structured, reliable music data 🛠️ Integrate music understanding directly into products and platforms via API 🧬 Detect AI-generated music and better understand AI involvement in catalogs 📊 Where we’re at • 55M+ songs analyzed with Cyanite • Trusted by 200+ music companies globally • 250,000+ artists and creators have used our technology • Used by companies including Warner Chappell, BeatStars, Epidemic Sound, LANDR, and APM Music 🧠 The team behind the technology Cyanite was founded in 2019 out of Popakademie Baden-Württemberg. Today, our team combines experience from Spotify, Sony AI, Fraunhofer, WhoSampled, and Twilio with research backgrounds from UPF Barcelona and TU Berlin, bringing a music-industry background together with deep expertise in technology and music information retrieval. 💡 What drives us We believe every piece of music deserves the chance to be understood and discovered based on how it actually sounds. By analyzing the audio itself, Cyanite helps people and technology work with music in more meaningful ways. Beyond commercial applications, Cyanite for Innovators supports educational and non-profit projects using music technology for research, culture, and innovation.

Founded 2019