{"type":"video","version":"1.0","html":"<iframe src=\"https://www.loom.com/embed/4aff607d2391487e9284bf67cfc3a0c4\" frameborder=\"0\" width=\"1838\" height=\"1378\" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>","height":1378,"width":1838,"provider_name":"Loom","provider_url":"https://www.loom.com","thumbnail_height":1378,"thumbnail_width":1838,"thumbnail_url":"https://cdn.loom.com/sessions/thumbnails/4aff607d2391487e9284bf67cfc3a0c4-fc6ef075ad3347b5.gif","duration":485.243,"title":"Music Recommender Demo with API Caching 🎶","description":"I built a music recommendation flow that selects moods and genres from what is in the file, then ranks the top five matches with a match percentage like 59 percent and energy boosts based on relative values. I added a growing cache mode and an API mode to fetch song and artist info. When you give input like jazz or synth, it can also use artist similarity if cached, and you can skip or try another option. I found artist similarity worked variably, for example with Nirvana suggesting Fleetwood Mac. I did not ask viewers for any action."}