Edit Mind transcribes and analyzes your footage entirely on-device, faces, objects, on-screen text, so you can search your library in plain English, and the new Strava integration uses GPS data to match action-camera clips to the exact activity they're from. Ilias Haddad (@iliashaddad3), a three-year YouTuber with 80 uploads, built it after a $450 bill from Google's cloud video API for analyzing a couple of videos pushed him to go local instead.
🔥 Our Take: Clipto and Invenio already do local AI video search, matching Strava's GPS track to the exact clip it came from is the part nobody else built. Three years and 80 videos in, Ilias knows the actual bottleneck isn't finding good footage, it's remembering which ride it's even from. Bullish, because a feature this specific only gets built by someone who's personally dug through hours of unlabeled clips looking for one climb.