Inspiration
POV and sports creators shoot hours of GoPro, Insta360, DJI and phone footage and store it on a NAS. Finding the few good seconds means scrubbing every clip by hand, so most of it is never used. Camera apps find highlights in their own brand's recent footage, but nothing searches the whole library where it already lives.
What it does
POV Moment Finder is a UGREEN AI NAS Skill. You point it at a folder and ask in plain English, for example "every jump I landed in 2025, under 60 seconds." It returns a ranked list of moments (file, start, end, and a one-line reason). After you approve the list, it writes a cut list (EDL) you can import into DaVinci Resolve or Premiere. It never uploads footage and never moves or deletes source files.
How we built it
Round 1 is the design: SKILL.md (goal, inputs, supported CLI capabilities, an indexer, ranker and presenter workflow, and completion and failure handling) plus a README and the proposal. It uses only Files and Video / Music content search from the UGREEN Skill Developer Package. Meta AI and Claude helped draft and edit the text.
What's next
The ranking itself, packaging against the Developer Package, the EDL export, and testing on a real NAS with real POV footage if selected as a Finalist.
Built With
- claude
- markdown