Florian Wenzel

Clean edit 🔥 Gave it a quick remix in Mirelo.ai to test sound fidelity - added SFX in one click, fully auto-synced and generated from the visuals frame by frame. So fun to experiment with!

🎧 Add sound to your videos — with 1 API call to Mirelo. All sound was created and synced by Mirelo. No prompts, no manual editing. It works on stylized, animated, or synthetic content - like this.

🎧 Watch this scene lock into rhythm with Mirelo. Mirelo creates in-scene audio that grooves with your visuals. Music and SFX that match your video frame by frame. Perfectly timed in seconds.

🎧Sound is a story of its own. Mirelo helps you shape it … with sound that matches your visuals, frame by frame.

🎵 Watch how to create perfectly synced sounds with Mirelo in seconds. Mirelo brings videos to life with in-scene audio that matches your visuals frame by frame. Layer multiple effects with precise timing. True to the moment.

🎧 Watch this scene lift off with sound by Mirelo Mirelo creates in-scene audio that matches your visuals. Always on cue. Mirelo: See it - Hear it - Instantly.

🤍Discover how Mirelo transforms videos with synchronised audio. Watch this scene move from day to night — and the sound transforming with it. Music shapes the look and feel of your video. Mirelo is there to make it perfect. Mirelo: See it - Hear it - Instantly.

New paper on the generalization of Flow Matching arxiv.org/abs/2506.03719 🤯 Why does flow matching generalize? Did you know that the flow matching target you're trying to learn **can only generate training points**? with @Qu3ntinB, Anne Gagneux & Rémi Emonet 👇👇👇

🎶Watch how to create perfectly synced sounds with Mirelo in seconds Mirelo brings videos to life with in-scene audio that matches your visuals frame by frame. Layer multiple effects with precise timing to craft immersive stories that captivate your audience.

🎶 Every moment has a sound. Mirelo finds it. Watch this little postcard come to life with sound. Mirelo creates in-scene audio that matches your visuals. Always on cue.

2 weeks ago, our co-founder & CTO Florian spoke at the @dida_ML Conference on Generating #Music & #Sound with Generative Foundation Models. Inspiring talks, amazing people, & great vibes all around. Thanks for having us! 👉 Try #Mirelo now & show us what you create! #MusicTech

Wohoo, or first product has launched in beta. Give it a try. Excited to see what you will do with it! :-)

🚀 Mirelo's beta is live! Creates perfectly synced sound and music for your videos! ✨ Try the beta at mirelo.ai 🎬 For creators and VideoGen artists 💬 Send [email protected] 📢 Share & tag us. We’re all ears! #Mirelo #GenAI #AudioAI #VideoGen #SFX #MusicAI

We are hiring at Mirelo. Work with us. :-) linkedin.com/jobs/view/4191…

Thank you @vincefort @markvanderwilk @EmtiyazKhan for organising this great @dagstuhl workshop! dagstuhl.de/24461

Also checking the space where apparently the cool kids hang out these days.

We are looking for research scientists/engineers to extend our team at Mirelo AI. If you're excited about building cutting edge foundation models for music and audio, join us! :-) linkedin.com/jobs/view/4048…

We have recently founded Mirelo AI building cutting-edge foundational models for music and audio. We are hiring research scientists/engineers. Join us! :-) linkedin.com/jobs/view/3837…

Super cool work from Max doing an internship with my old team at AWS. In the paper, he shows: Despite the fuzz, currently, diffusion models are not yet efficiently leveraged for augmentation. We tested many different approaches, but none could beat a simple retrieval baseline.

🎊Extremely honored to share that our paper on multimodal model robustness has been accepted as the 1st paper for the Journal of Data-centric Machine Learning Research @DMLRJournal With @yizhu59 @sxjscience @flwenz @mli65 #Multimodal #Robustness #DistributionShift