For a short you already posted. Send the video plus a screenshot of its retention graph (or your average % viewed) and get the curve redrawn with the drop-offs marked, the frame on screen at each drop, every leak ranked, a cut list with the new target length, and a verdict printed next to its threshold.
Upload one video and ffmpeg pulls candidate frames by scene change and even sampling, then scores them on sharpness, contrast, faces and colorfulness. A 2 to 6 word hook phrase goes on the chosen frame in 3 layouts (side panel, bottom stripe, top banner), giving 1280x720 JPG thumbnails A/B/C (under 2MB, YouTube spec). A comparison sheet shows full size next to the 168x94 mobile feed size, so you do not pick from the zoomed view only. Runs locally, no API key.
Diagnose why the first 6 seconds of your short-form video loses viewers — get a second-by-second evidence timeline, up to 3 named findings, 3 structurally different replacement hooks, and a timecoded edit table you can cut today.
Send your short video and get its first six seconds read frame by frame: what is on screen and heard each second, where the picture freezes or the sound drops, a risk level per second with the reason, and three replacement openings to film. One report image plus the findings, for TikTok, Reels and YouTube Shorts.
Paste each new Shorts topic and get a feed-ready title, a 3-second hook, and marked retention beats at 15s and 30s so the script keeps paying off. It never claims YouTube Analytics or view-count data.
Predict second-by-second viewer retention for the first six seconds of a short-form video hook, then provide concrete opening fixes.
Create branded PNG and animated MP4 batches from approved local visuals and copy. Reuse brand settings across three templates and three aspect ratios, review previews, then rerender only changed final jobs. Complete English and Traditional Chinese workflows. Terminal setup, Python, FFmpeg and your licensed font are required.
Plan a conversion-focused seven-slot Amazon listing image gallery and generate compliant image briefs locally.
Check your .srt or .vtt captions before you post. An offline script finds exact cues with reading speed, line length, overlaps, timing errors, broken characters and script mismatches, fixes the mechanical problems without changing a word, and re-checks. It cannot hear audio or test platform acceptance. Download package: bring your own AI agent and Python 3.