I built an edit-system demo for an executive interview episode on a placeholder transcript. It specifies the reusable project: bins for Footage, Audio, Graphics and Final Timelines with 20 sub-bins, four timelines, three render presets and a naming rule that rejects overwrites. It designs the graphics the macros carry (lower third, caption style, call-to-action overlay, chapter card) in 16:9 and 9:16, each placed inside computed safe areas, with the vertical keeping platform interface zones clear. SRT captions are written for both formats, two lines at most and at most 17 characters a second: 33 and 53 cues. It reviews the posted brand colors by WCAG contrast (white on the blue 3.30 to 1, on the purple 7.12 to 1, on the green 2.86 to 1) and turns the numbers into six recommendations. 14 passages are scored on four questions to find what lands, a 79.6 second reel is assembled in five beats with 5 clips of 15 to 30 seconds, and the volunteer walkthrough script is written. The build asserts every timing, safe area and caption rule before writing anything.